777 Heathrow

[MUSKEG]

Ok so we are having accidents that are pilot induced. Does anyone have any idea how many accidents pilots have prevented by taking over from the computor? Of course no one hears about them, thats our job. You guys are gullible if you think that they will ever fly 500 people arround and not have crew on board. Sure the airforce has pilotless aircraft. Why? because they aren't hauling people. If you lose one it's just metal, expensive yes, but painful no. Hey everyone I have ocean front property for sale in paraguay. The public will not support the idea because our comps crash at home.

[hairdo]

even if you remove pilots from the aircraft, you cannot remove human error. until computers program themselves, there will always be humans programming them, and as we all know, humans are prone to make errors...

[stmymdy]

like its been said before, it all comes down to adaptability.
a computer simply relies on its programming. it cant adapt to unusual situations like people can.

[CID]

Currently, humans and computers combine to give us the safety we enjoy when flying in modern airliners.

Ok so we are having accidents that are pilot induced. Does anyone have any idea how many accidents pilots have prevented by taking over from the computor?

If the computers work and the pilots do their jobs properly, safety is the probable outcome. If either deviate, safety can be compromised.

like its been said before, it all comes down to adaptability.
a computer simply relies on its programming. it cant adapt to unusual situations like people can.

Sometimes the "unusual situations" are induced by pilot error. Variability is reduced (if not elimiated) by use of technology. Pilots can be easily replaced with the use of automation and redundancy. The public however isn't going to accept pilotless airplanes any time soon though.

With proper programming and redundancy, the rate of computer error and inability to deal with variables outperform 2 human pilots if you consider the cause of all accidents. Ultimately, with such a system, you may occasionally see an accident that may have been prevented with a human in the seat but you'd see less accidents caused by innappropriate operation or failure to operate the airplane in accordance with the published procedures and limitations.

An extreme example would be the Egypt Air 767 that was forced down by the deliberate actions of the copilot. A computer (or 3 or 4 of them) if programmed properly (as verified by appropriate certification flight testing) wouldn't base any of it's flight control commands on its religion.

Something to think about.

[Dust Devil]

like its been said before, it all comes down to adaptability.
a computer simply relies on its programming. it cant adapt to unusual situations like people can.

Autopilots adapt all the time. That's what they do

[Troubleshot]

like its been said before, it all comes down to adaptability.
a computer simply relies on its programming. it cant adapt to unusual situations like people can.

Autopilots adapt all the time. That's what they do

yeah but they don't gain new information and make up their minds on what to do based on situations, they look at logic gates to determine the next course of action.

[Mig29]

there goes this BA777 thread too.....oh well

can we make this discussion automated as well somehow??

[flyinhigh]

stopsquawk

Back to the original subject: Just wondering, if the cause for this engine rundown was some yet undetermined mysterious software or fuel system problem, wouldn't the fleet have been grounded as a precaution? or maybe new procedures been immediately put in place to keep it from happening again? Were 777 pilots directed to change procedures or take any other precautions following the accident? I've long suspected that the exact causes of the crash was immediately determined, or were known very shortly following the accident. Otherwise there would have been some reaction from the industry, like there was during SAS Q400 landing gear problems last year.

So does anyone have a answer for this question?

[small penguin]

Some of you cranky pilots need to go play some video games. FarCry comes to mind. Then come back and tell me the enemy AI didnt adapt to you and kick your ass.

Computer AI has come a long way, and will go a long way. So has computer hardware. Is it perfect? Not at all! Are you (ATPL pilots) perfect? Not at all either!

[valvelifter]

Guys, I work with software every day as a tester/troubleshooter. Software itself with 'proper' testing is usually pretty stable. When software is released to some testing org, a series of test cases are created and testing follows this plan. When a fault arises, the org usually uses some sort of trouble report handling procedure where the faults/fixes are fed back to the design organization, who fixes them and releases upgraded s/w, and the cycle continues until it is released, deployed, and maintained. So if testing is done correctly, then the software is usually pretty stable. That is, assuming that ALL the test cases were executed. Sometimes, due to timing, cost, or market conditions, certain test cases are analyzed, a risk assessment is done on skipping them, and a decision whether or not to drop these test cases is made. Usually, there are a number of TCs that are dropped as you can't usually test everything with the budget that is allocated.

The problem is that no testing org. can test ALL combinations of conditions. Testing assumes to some extent that we live in a perfect world.

If you have 26 variables into a piece of s/w for instance, you cannot test all possible combinations of variables because you would be testing the s/w forever. And multiply this for every process in the system. Then you have to test interworking between all the processes, when they are in different states, etc... It's just impossible to test everything.

Just as an example, suppose an engine malfunction happened at exactly the same time the auto-throttles are disarmed. Or if a rare software fault that would normally be transparent to the user (i.e. pilot) happened while the gear is dropped, and this escalated into a more serious fault where there is corruption in the code....the possibilities are endless. I work in mobile communications and the severity of a fault is, well, the customer might get a dropped call. Not too bad, and even for this type of industry, where lives are not directly at stake, the equipment is monitored 24 hours a day and support people are on call (and called very often by the way).

Now imagine an aircraft with x souls on board? I can't imagine that we will get to the point where no one is required on board. At least not with the current technology we have. Especially with the fact that many companies are now outsourcing these design / testing jobs to third party companies, sometimes located in countries where they are paid in one day as much as you would make in an hour. And from experience, it is usually not cost-effective to do this, although the guys with the ties usually go this way b/c it looks good in a powerpoint presentation.

Having 'monitors' on board, ready to save the day in case S*&t hits the fan. Well, where are the cost savings there if you have to still pay a pilot or two to stay on board, and also beef up support procedures on the ground to handle every type of fault, and fast?

[teacher]

Saw this the other day, thought I'd add it to the mix. Interesting about the fuel pump wear.

Fuel feed interruption still in the frame for BA 777 accident, says AAIB

By David Learmount

Partial interruption to the fuel flow still appears to be a potential cause of the British Airways Boeing 777 crash-landing at London Heathrow on 17 January, according to the latest UK Air Accidents Investigation Branch (AAIB) bulletin. The agency is careful not to state it as a cause, but concludes the interim report by saying that "a comprehensive examination and analysis is to be conducted of the entire aircraft and engine fuel system, including the modelling of fuel flows and taking account of the environmental and aerodynamic effects".

The 777 (G-YMMM) - which has been declared beyond economical repair - was descending through 720ft (220m) on final approach with the runway in sight when autothrottle, and then the crew, demanded an increase in thrust, but they only got a temporary response, following which the power reduced to the same level. The report says: "By 200ft the airspeed had reduced to about 108kt (200km/h). The autopilot disconnected at approximately 175ft, the aircraft descended rapidly and its landing gear made contact with the ground some 1,000ft short of the paved runway surface." At 108kt the aircraft was within a couple of knots of stalling speed.

The AAIB says it has tested all the electric and electronic engine and fuel system controls and found them to be working perfectly, and there is no indication in the flight data recorder of any such malfunction during the flight or the event itself. Flight data shows that, at the moment of thrust reduction, the right-hand engine's electronic control system was responding correctly to a reduction in fuel flow to the right engine. This was followed by a similar response from the control system in the left engine when fuel flow to that powerplant also diminished.

The agency remarks, however: "Detailed examination of both the left and right engine high pressure fuel pumps revealed signs of abnormal cavitation on the pressure-side bearings and the outlet ports. This could be indicative of either a restriction in the fuel supply to the pumps or excessive aeration of the fuel. The manufacturer assessed both pumps as still being capable of delivering full fuel flow." The AAIB does not say whether it associates this evidence with the accident cause or not, but adds: "Investigations are now under way in an attempt to replicate the damage seen to the engine high pressure fuel pumps and to match this to the data recorded on the accident flight."

"There was no evidence of a mechanical defect or ingestion of birds or ice," says the AAIB.

http://www.flightglobal.com/articles/20 ... -says.html