Slashdot Mirror
Failed Avionics a Possible Cause of BA038 Crash
Muhammar writes "As you may have heard by now, both engines of the Boeing 777 aircraft flight BA038 suddenly cut off without warning at very low altitude and low speed during autopilot-assisted landing at Heathrow. A prompt reaction of the pilots prevented the stall and saved all lives aboard. The crash landing short of the runway tore off the landing gear on impact, and the fuselage plowed a long, deep gouge in the grass. With the investigation ongoing, the available information points to an electronic control problem as the most likely cause of the sudden engine power loss."
A bit of FUD here I think - unless I read TFA wrong, the entire thing is under investigation and no one is saying anything for at least a month. The autopilot apparently sensed the need for more thrust and warned the pilots of this. It might be premature to say that a software problem is the likely cause of failure...
"As the intrepid kobold companion continues his journey, he begins to wonder... if priests raises dead, why anybody die?
The pilots then manually increased throttle - to no avail.
For both engines to malfunction like this at the same time greatly seems to point to a fuel delivery problem.
This does not necessarily mean "running out of gas" - as a plane like this has multiple tanks, valves and pumps, all of which can be configured multiple different ways - which change during the flight.
A simplistic example: they could have been running both engines off one tank - which went dry - though another was full - or both engines were being fed from a common fuel pump which failed, etc. These things *shouldn't* happen - but the investigation will tell...
Let's just wait for the official forensics rather than patched together rumours shall we?
AT&ROFLMAO
The word "hero" is thrown around a lot these days...
I believe what they meant, was that the pilots realized that things were going wrong, and the "normal" reaction would be to add thrust. When they realized that they couldn't add thrust, that this would result in loosing airspeed, entering a stall, and crashing
So they realized that an alternative was to lower their angle-of-attack, preventing the stall, and maintaining a bit of airspeed. This would have the unfortunate side affect of landing well-short of the runway (and perhaps airport) and destroying the aircraft - but given the information available - was a bad - but the best alternative
So they implicitly decided the best course of action was to glide the airplane and ditch it in a field - not a decision that would have exactly won them any praise had they read the situation wrong - but it saved everyone
To my mind, if you manage to get 300 tonnes of falling metal out of the sky and on the deck with nothing worse than a broken leg, you've done something right.
[FUCK BETA]
Not a commercial aircraft airframe and powerplant mechanic, but I was a senior avionics technician for many years dealing with corporate and private jets.
What I've read is that the pilots observed a relatively gradual loss of power symmetrically on both engines.
Interesting. Do you have a link to the source for that? Not that I doubt you, just curious to parse it myself.
This tells me that I can rule out engine problems with FADEC and fuel.
FADEC, possibly, but fuel? It's quite possible there was either water or crud in the fuel, especially since the aircraft almost certainly took on fuel in China, and China seems to have had problems of late with products being adulterated in some form. The crud could cause blockages in the filters from the tank(s). The water would cause an increasingly-diluted fuel mixture to enter the engines as the level dropped which might also cause the gradual loss of power.
The two most-likely culprits I would examine first are the discrete devices at either end of the control path that send the data and receive it at the other end, and the cables and connectors used to transmit the data.
The next point I'd check would be the power supply that powers the electrical actuators that physically move the actual throttles in each engine. This supply would be separate from the power used for the electronics, as it would be a relatively high-current source. This might also be caused by cabling/connector problems.
Aircraft tend to have many problems with cabling due to high vibration and multiple pinch-points and stress and vibration/abrasion at support points, as well as contact problems at connectors.
Another very major problem is human error. In many cases the turn-to-lock type connectors are in very tight spaces, sometimes so much so that it may only be visible by a small mirror and flashlight held by the tech while he may be laying on his back or nearly standing on his head. I had a whole set of strange-looking pliers of different lengths and weird angles with curved padded jaws for just this purpose in my tool box, along with small hand-held extend-able flexible-tubing-mounted inspection mirrors and flashlights with the head on flexible tubing as well.
It can be very hard to tell, given the above circumstances, if the locking sleeve on these aircraft instrumentation connectors had been twisted far enough to complete the lock. It doesn't take much imagination to see what could happen given time, vibration, and G-forces.
Of course, these are just my rough guesses, and I don't have enough information to really make any informed statements.
Cheers!
Strat
Progressivism (aka US 'Liberalism'): Ideas so good they need a police/surveillance-state to enforce.
Given that the plane is heavily instrumented, available, and didn't burn, this should be a simpler case to examine. Hopefully, a lot can be learned. At least more than if it crashed and burned in a jungle, or into the ocean.
In principle, the airplane could have been landed on the runway without damage, if the right variables had come together -- but low and slow, in a big heavy airplane, with full flaps and no power, you're pretty well boxed in. I don't think they could have done better.
rj
Posting anon for obvious reasons.
:)
I work in the avionics industry and this was exactly my thought as well. These systems are becoming much more complex than you would expect embedded software to be. Several address spaces and over a dozen threads is fairly normal with most newer systems.
Typically the safety critical industry likes to tout itself as being better designed than other software because it conforms to various standards, particularly do178b. At their core, these standards basically say you need to have processes that everyone understands in place when you design your software and you need have documentation that shows you tested all the different elements of functionality. The testing may be fairly rigorous depending on who is doing it, but at the end of the day they arent doing much that microsoft/oracle/your favorite well known software vendor doesnt do. (although I am sure that many here beleive that ms doesnt test its software)
That sounds really dumb. Tools that can verify that software matches the specifications 100% in every case under every condition? For anything but the most rudimentary code I seriously doubt that. There was a relatively recent incident where a 777 gave warnings that it was going too fast and too slow, both at the same time. Attributed IIRC to a failed sensor and software not programmed to handle the error correctly. That blows the 100% software verification test suite right out of the water. If they really adopted that methodology they probably did it for economic reasons rather than safety.
"This is your automated pilot speaking. Sit back and enjoy your flight with us this afternoon on the first completely automatic airliner. Nothing can go wrong... go wrong... go wrong... go wrong."