Airline Crash Due to Crew Errors

[FLYBOYJ]

So just a process check here. Can you restate the point of this thread again? Is it to discuss whether UAS' are inherently more safe than manned aircraft?

Well if it goes back to the first post, the two examples Sys gave were VERY poor.

 

[Matt308]

Would you allow your family to fly in an unmanned UAV right now?

Even if his answer is yes, there is not a civil aviation authority on the planet that would allow him to.

I'm wondering if the point get's lost in the argument. Everyone seems to be in violent agreement that UAS' are a reality today, in the near future and long-term. I'm just struggling with the point of this thread.

 

[mkloby]

I'd just like to add my 2 cents, as I fly an aircraft that is highly automated on a daily basis.

Automation is a good thing, and can significantly reduce pilot workload, but it has to be used correctly.

Syscom - it is no secret that human error plays a part in most mishaps. Human error plays a factor in approximately 80-85% of mishaps, if I recall correctly. There will still be human factors with a human operating a UAV, no doubt.

Also, you need to be VERY careful about designing aircraft that restrict the pilots freedom of action. Some automation is a great thing, but some can be very bad. You need to understand that aircraft systems often do not work as the engineers imagine. Parts fail in high performance aircraft all the time. It is not a perfect world. When aircraft systems are tied together to large extent through computers, software, data buses, etc., small part failures can have large consequences.

 

[syscom3]

So just a process check here. Can you restate the point of this thread again? Is it to discuss whether UAS' are inherently more safe than manned aircraft?

Yes.

 

[syscom3]

Syscom - it is no secret that human error plays a part in most mishaps. Human error plays a factor in approximately 80-85% of mishaps, if I recall correctly. There will still be human factors with a human operating a UAV, no doubt

And if the human element is reduced even further, then the number of human induced accidents goes even lower. And with current software and processors, theres nothing stopping a UAV or remotely piloted vehicle from being even safer, by preventing stupid things from happening in the first place.

MK, think about the human limitations on processing and correctly coming up with a solution when faced under extreme stress, multiple failures and you only have a few seconds to "fix" it, and if not successfull, you crash and die.

UAV's and computer controlled cockpits dont have those limitations, thus by default, they are safer.

Parts fail in high performance aircraft all the time. It is not a perfect world. When aircraft systems are tied together to large extent through computers, software, data buses, etc., small part failures can have large consequences.

Yes, Ive never doubted that. I also am a strong believer in testing in the real world, so when automation takes place, its not an ad hoc prototypical solution.

Also, you have to look at the accident statistics as a whole. If a completely automated system has a 1 in a million chance of complete failure, and we will lose an AC once every year .... but .... the automation feature means we have also prevented 3 or 4 crashes in the same time period, then things are safer.

 

[syscom3]

Obvious? Not in a million years. And statements that application of sound human factors is not "complicated" is boundlessly naive. Surely can't expect such statements to go unchallenged, sys. Properly applied human factors analyses are immeasurably complex and in many cases very subjective. And identifying system engineering solutions as means of mitigating human factors risks only adds to the complexity of the hardware and software development.

Heres some simple code for you:

"Do not allow landing gear to be retracted untill altitude "X" is reached".

"Do not allow aircraft to enter a bank exceeding X degrees at Y airspeed so as to prevent stalling or exceed allowable gee loads".

"Do not allow AC to touchdown for landing if air speed exceeds Z MPH and runway length is too short".

Are you telling me this is complex?

BTW, the B2 and F16 are unstable aircraft to fly. Dont you suppose that the avionics is whats allowing the plane to fly to begin with?

 

[syscom3]

Well if it goes back to the first post, the two examples Sys gave were VERY poor.

Why is that? tell me why they are poor. In fact, they're textbook examples of human failure allowing an AC to crash.

And sorry, saying that modern crew flight procedures and discipline wont let that happen again is incorrect. As long as people are people, sooner or later cockpit discipline will become lax and flight procedures "over looked" from time to time.

Just like that Russian captain who let his kids fly his airbus. Who would have thought ......

 

[FLYBOYJ]

Why is that? tell me why they are poor. In fact, they're textbook examples of human failure allowing an AC to crash.

And a UAV operator can do the same thing while troubleshooting a problem. The L-1011 is texbook for training scenarios and would also apply to UAVs

And sorry, saying that modern crew flight procedures and discipline wont let that happen again is incorrect. As long as people are people, sooner or later cockpit discipline will become lax and flight procedures "over looked" from time to time.

And the same thing applies to any machinery. Add a computer to run it and you have rigidity in the decision making ability

Just like that Russian captain who let his kids fly his airbus. Who would have thought ......

Unrelated to this discussion....

 

[FLYBOYJ]

Heres some simple code for you:

"Do not allow landing gear to be retracted untill altitude "X" is reached".

And you cannot gain immediate positive climb, something desirable in jet aircraft

"Do not allow aircraft to enter a bank exceeding X degrees at Y airspeed so as to prevent stalling or exceed allowable gee loads".

And now you're limiting the manevability of the aircraft

"Do not allow AC to touchdown for landing if air speed exceeds Z MPH and runway length is too short".

And how does the UAV determine runway length? What if you're operating on a dirt strip? If you have ice on the airvehicle you have to land at a higher than normal airspeed.

Are you telling me this is complex?

It's not but then you limit the capability of the unit. What I just described has to be determined by a human

BTW, the B2 and F16 are unstable aircraft to fly. Dont you suppose that the avionics is whats allowing the plane to fly to begin with?

They are, but the systems that allow those aircraft to fly are synthesized to allow human decision making in most of the flight envelope.

 

[syscom3]

Here's another example where pilot error went to the extreme and cockpit automation would have prevented this from happening.

B-52 crash at Fairchild Air Force Base occurred on June 24, 1994, killing the four crew members of a United States Air Force (USAF) B-52 Stratofortress named Czar 52[1] during an airshow practice flight. In the crash, Bud Holland, who was the command pilot of the aircraft based at Fairchild Air Force Base, flew the aircraft beyond its operational limits and lost control. As a result, the aircraft stalled, hit the ground, and was destroyed.

1994 Fairchild Air Force Base B-52 crash - Wikipedia, the free encyclopedia


Here was complete failure of pilot dicipline and flight deck dicipline. And if the AC had a computer flying the plane in which prohibited or unsafe maneuvers were prevented from happening, then this would not have occurred. And if the computer is flying the plane, why should there be a pilot?

 

[FLYBOYJ]

Here's another example where pilot error went to the extreme and cockpit automation would have prevented this from happening.

B-52 crash at Fairchild Air Force Base occurred on June 24, 1994, killing the four crew members of a United States Air Force (USAF) B-52 Stratofortress named Czar 52[1] during an airshow practice flight. In the crash, Bud Holland, who was the command pilot of the aircraft based at Fairchild Air Force Base, flew the aircraft beyond its operational limits and lost control. As a result, the aircraft stalled, hit the ground, and was destroyed.


View: https://www.youtube.com/watch?v=-kHa3WNerjU


Plenty of computers working that one Sys...

Some comments...

Don't be surprised if the Airbus fly by wire computers didn't put a perfectly good airplane in the water.

Google Airbus A320 Crash at the Paris Airshow in 1998. Watch the video of an airbus A320 crash into a forest because the computers wouldn't allow a power increase following a low pass. The computers wouldn't allow a power increase because they determined that the airspeed was too low for the increase requested so the computers didn't give them any. Pushing the throttles forward in a Airbus does nothing more than request a power increase from the computer. If the computer doesn't like all the airplane and engine parameters you don't get a power increase. Airbus blamed the dead crew since they couldn't defend themselves. A Boeing would still be flying.

How Airbus nearly killed 155 people according to an idiot - Plane Talking

 

[Matt308]

You stole my thunder Joe. I was gonna cite the A320 crash. To this day the flight control laws of the Airbus aircraft are subject to heated debate. Boeing has the exact opposite philosophy, "You wanna bend the airframe, that is the pilot's call... not the FCC". My quote... not Boeing's.

 

[FLYBOYJ]

You stole my thunder Joe. I was gonna cite the A320 crash. To this day the flight control laws of the Airbus aircraft are subject to heated debate. Boeing has the exact opposite philosophy, "You wanna bend the airframe, that is the pilot's call... not the FCC". My quote... not Boeing's.

Here's more on this - on that site there are a few guys slamming the submitter and he does seem to be "Boeing Biased."

AirDisaster.Com: Investigations: Air France 296

 

[DerAdlerIstGelandet]

There was German company or govt research institute (I would like to say BMW but I am not sure) that demonstrated "autopiloted" cars used on both highway and city driving and it worked marvelously. Only problem was it was too expensive for consumers. But it worked. I also know there are some semi-automatic engine/throttle/steering controls that work like ABS's do right now .... prevent the car from becoming uncontrollable or unsteerable.

That fact that you actually believe that computers can do the job more safely than a human is very troublesome. You worry me...

The AC crashes. But the failure points for an automated cockpit are far fewer than manned cockpits, simply because the human element is taken out.

A simple equation for you:
w = number of crashes soley due to pilot error
x = number of crashes due to a combination of human error and mechanical failure
y= number of crashes due to mechanical failure only.
z = number of crashes due to events beyond the control of the AC or pilot (like bird strikes)

N = w+x+y+z.

Now if we eliminate variable w, already we are < N

and factor in part of x, in which pilot error made a bad situation worse, or was the primary cause of the crash, then "N" is even lower.

Would a UAV that's automated to a high degree, with properly tested software to accommodate failures, could still crash? Yes, statistically it will still happen. But it is still obvious that the UAV would be inherently safer than a manned AC. And even if there is an external operator flying the UAV, there's nothing complicated in writing "do not do or allow" software to be written to prevent accidents, or operating the AC out of its limits.

No you are looking at this in black and white. There is a lot more to the picture than that.

Aircraft already fly themselves for the most part. Why are pilots needed? For when things go wrong. A computer can never replace a human in that area.

Do accidents happen because of Human area? Of course, but that does not make airline travel any less safe.

Utter falsehood.

He obviously has never worked with government computer systems...

Heres some simple code for you:

"Do not allow landing gear to be retracted untill altitude "X" is reached".

"Do not allow aircraft to enter a bank exceeding X degrees at Y airspeed so as to prevent stalling or exceed allowable gee loads".

"Do not allow AC to touchdown for landing if air speed exceeds Z MPH and runway length is too short".

Are you telling me this is complex?

BTW, the B2 and F16 are unstable aircraft to fly. Dont you suppose that the avionics is whats allowing the plane to fly to begin with?

Sys that stuff is already built into onboard computers. That however does take the need away from a pilot.

Here's another example where pilot error went to the extreme and cockpit automation would have prevented this from happening.

B-52 crash at Fairchild Air Force Base occurred on June 24, 1994, killing the four crew members of a United States Air Force (USAF) B-52 Stratofortress named Czar 52[1] during an airshow practice flight. In the crash, Bud Holland, who was the command pilot of the aircraft based at Fairchild Air Force Base, flew the aircraft beyond its operational limits and lost control. As a result, the aircraft stalled, hit the ground, and was destroyed.

1994 Fairchild Air Force Base B-52 crash - Wikipedia, the free encyclopedia


Here was complete failure of pilot dicipline and flight deck dicipline. And if the AC had a computer flying the plane in which prohibited or unsafe maneuvers were prevented from happening, then this would not have occurred. And if the computer is flying the plane, why should there be a pilot?

I can do the same thing. Here is a perfect example of human error.

Mother Drives Car Into Apartment Building - cbs2.com

Who cares, my point?. You are doing nothing but grabbing for air here, and millions of people still drive cars...

 

[DerAdlerIstGelandet]

I think everyone agrees and is well aware that automation in aircraft is a good thing. It really does help the pilot by taking some of the work load. In the end though fact is fact and something sys fails to realize is that you can not completely replace a human in the cockpit (at least not at this time).

 

[Matt308]

Heres some simple code for you:

"Do not allow landing gear to be retracted untill altitude "X" is reached".

"Do not allow aircraft to enter a bank exceeding X degrees at Y airspeed so as to prevent stalling or exceed allowable gee loads".

"Do not allow AC to touchdown for landing if air speed exceeds Z MPH and runway length is too short".

Are you telling me this is complex?

Yes I am. Integrated Flight Control Computers with airframe flight limitations is a very significant engineering problem. Now your last example couples not only flight control algorithms, but also introduces dynamic modifications based upon aeronautical information service (AIS) parameters. These AIS parameters are are identified in ICAO Annex 11, but are in no way standardized for civil aerodromes (airports). You have yet again made an engineering oversimplification that your "flowchart" you are working would say...

 

[FLYBOYJ]

UAV reliability

great info!

 

[Matt308]

Phhhttt... more facts to get into the way of a good argument.

 

[FLYBOYJ]

 

[syscom3]

Flyboy, explain this one .... Perris Valley Airport. 1984. Drunk pilot crashed into a DC3. I missed this crash by an hour, and had just jumped out of the DC3. Pilot error all the way. An automated cockpit wouldn't have allowed this to happen. Drunk pilots is an issue automated cockpits dont have to deal with. All the laws on the books didnt stop this pilot from some serious lapses in judgement.

Just like that Russian captain who let his kids fly his airbus. Who would have thought ......
Unrelated to this discussion....

So I take it you dont have an answer to pilots committing lapses in judgment? What about cockpit discipline and prcedure's? I suppose the pilot knew all about those and violated them anyways. Automated cockpits dont have these issues. They're not influenced by human frailties.

As for the A320 crash .... excellent point. And it obviously shows that thourough testing must be done to prevent errors like this from happening. But then ...... if this type of accident only happens once per decade, and the flight computers have prevented several crashes before then, then again, things are safer.

Flyboy, as for your comment about decision charts .... I am not trying to insult your intelligence, but these charts are used extensively in flight manuals and as the basis for automated flight controls, and have been for decades. So I apologize if I misread your statement or you were not clear in what youre saying.

Deradler ..

"Sys that stuff is already built into onboard computers. That however does take the need away from a pilot."

I am pointing out that some pilot errors are just so stupid, they are or can be preventable by the on board avioncs. There is nothing complicated for a autopilot to refuse to retract the landing gear if they're still on the ground, or refuse to shutdown an engine on takeoff if there is no detectable problems, nor prevent a pilot from putting the AC into a maneuver where it will fail catastrophically or go out of control.