Taiwan Plane Crash

[Simple Jack]

Feathered? You mean like bird strike?

http://www.telegraph.co.uk/news/worldnews/northamerica/usa/8581244/Major-bird-strike-incidents.html

No, and forgive me because I'm terrible at explaining concepts in an understandable way but:
On a twin (or more) engined propellor driven aeroplane (or most single engine turboprops), whether they be piston driven or turbine driven (turboprop, the type in question) with variable pitch propellors (which is almost all of them), you have the ability to rotate the propellor 90 degrees so it sits streamlined in the airflow. This is critical because a 'windmilling' propellor causes an absolute shitload of drag. A failed engine with the prop not feathered will windmill in the airflow causing this drag. Many types of twin prop aircraft just will not climb with the failed engine windmilling.

I was saying that one of the props looks to be at a vastly different pitch to the other indicating one failed engine, however it's almost impossible to tell from a photo and it could easily be a visual illusion.

 

[Geelong_Sicko]

No, and forgive me because I'm terrible at explaining concepts in an understandable way but:
On a twin (or more) engined propellor driven aeroplane (or most single engine turboprops), whether they be piston driven or turbine driven (turboprop, the type in question) with variable pitch propellors (which is almost all of them), you have the ability to rotate the propellor 90 degrees so it sits streamlined in the airflow. This is critical because a 'windmilling' propellor causes an absolute shitload of drag. A failed engine with the prop not feathered will windmill in the airflow causing this drag. Many types of twin prop aircraft just will not climb with the failed engine windmilling.

I was saying that one of the props looks to be at a vastly different pitch to the other indicating one failed engine, however it's almost impossible to tell from a photo and it could easily be a visual illusion.

I think I get it! So when an engine works properly the propeller blades themselves can be tilted into the wind, whereas a failed engine can't rotate the blades at all, leading to excess drag and problems climbing?

 

[DrKrieger]

I think I get it! So when an engine works properly the propeller blades themselves can be tilted into the wind, whereas a failed engine can't rotate the blades at all, leading to excess drag and problems climbing?

Yep, pretty much!

Here's a video showing a prop being feathered for training, in the video the engine is shut down, but the prop continues to spin because of the air flowing over it (due to the aircraft still moving through the sky), so it is then feathered at around 0:06. Aerodynamically, the spinning prop is pretty much the same as having a solid circular disc the size of the prop attached to the front of the engine, which creates a tonne of drag. Feathering makes a huge difference.



It's extra important to get that reduced drag in a conventional twin engine aircraft, as the working engine at high power on one wing, plus the drag on the prop of the failed engine on the other wing will work together to try and twist the plane around in the direction of the failed engine. This is countered mainly with the rudder, but the effectiveness of the rudder is directly related to how much air is flowing over it (i.e. more speed = more effective rudder).

If the aircraft slows down too much, the rudder will have less directional authority than the working engine, and the pilot won't be able to stop the aircraft pulling around to the failed side unless they reduce power on the good engine, or find more speed (usually by descending to get some help from gravity). This speed is the Vmca referred to earlier.

This gives a pretty simple/good explanation of how it all works.

 

[Simple Jack]

I think I get it! So when an engine works properly the propeller blades themselves can be tilted into the wind, whereas a failed engine can't rotate the blades at all, leading to excess drag and problems climbing?

Not quite, when the engine is working they can be adjusted but stay within a normal 'facing the airflow' range, when the engine is failed it's essential to get them turned to the streamlined position because if they stay in the normal position the airflow keeps turning the blades (like a windmill) causing excessive drag.

Failure to promptly feather the failed engine has contributed to quite a few crashes.

 

[Simple Jack]

Are you an instructor DrKrieger ? I couldn't explain that stuff as well if you gave me a million goes at it.

 

[True__Believer]

Some great stuff there guys

 

[DrKrieger]

Not quite, when the engine is working they can be adjusted but stay within a normal 'facing the airflow' range, when the engine is failed it's essential to get them turned to the streamlined position because if they stay in the normal position the airflow keeps turning the blades (like a windmill) causing excessive drag.

Failure to promptly feather the failed engine has contributed to quite a few crashes.

Argh, missed that part!

Just to add on, the propeller pitch is normally hydraulically actuated (oil pressure etc), so as long as the prop is spinning, there should be enough pressure in the system from the pump to be able to feather it. (And the prop spinning is more or less the reason to feather it, if it's not feathered the airflow will keep it turning at pretty much all safe speeds).

In some turboprop designs the "neutral" position for the propeller is feathered, and the engine has to pump oil into the propeller system to generate the force required to keep it in the normal "flying" position, so lose the engine and you lose that pressure, and the prop defaults back to feathered. I'd be guessing, but the ATR props are normally feathered when shut down on the ground, so this could be one of those systems.

Also, have had another look at the video on a bigger screen, no doubt VMCA played a part, I do still think it looks an awful lot like a stall just from the way the left wing seems to let go and drop, and the aircraft looks like it rolls around the longitudinal access without a lot of yaw. Could be just the perspective though, or I could just be wrong, which certainly isn't unheard of.

 

[Simple Jack]

Also, have had another look at the video on a bigger screen, no doubt VMCA played a part, I do still think it looks an awful lot like a stall just from the way the left wing seems to let go and drop, and the aircraft looks like it rolls around the longitudinal access without a lot of yaw. Could be just the perspective though, or I could just be wrong, which certainly isn't unheard of.

I've been reliably informed by a mate who's current on the ATR that it doesn't have the roll authority to lower the wing that quickly. He reckons it's definitely stalled and dropping a wing.

 

[True__Believer]

I've been reliably informed by a mate who's current on the ATR that it doesn't have the roll authority to lower the wing that quickly. He reckons it's definitely stalled and dropping a wing.

So in the context of this flight, is this looking like a combination of vmca roll with a stall in the final moments? Or something else completely? Planes such as these are required to be able to keep climbing and at least maintain airspeed on only one engine don't they?

 

[peternorth]

have just seen footage of the plane clipping the taxi, and the subsequent phone call from driver to taiwans 911/000. wow.

here in the birthplace of the father of taiwan there is news on the tv via international news conduits. however, i have yet to be engaged in conversation with locals about this.

 

[rippersnipper]

To be fair, I feel easier when I'm in control of the plane

 

[True__Believer]

Initial reports trickling out that they had a problem with the right engine, but the pilots then proceeded to switch off the LEFT engine, which they were then unable to restart. Catastrophic error if true.

 

[Simple Jack]

So in the context of this flight, is this looking like a combination of vmca roll with a stall in the final moments? Or something else completely? Planes such as these are required to be able to keep climbing and at least maintain airspeed on only one engine don't they?

I watched it again. It's very hard to tell with videos from the ground but it seems like it's already stalled wings level before the wing drops. Once again you really need the data to tell.

Initial reports trickling out that they had a problem with the right engine, but the pilots then proceeded to switch off the LEFT engine, which they were then unable to restart. Catastrophic error if true.

Yeah, that'd do it.

 

[rdhopkins2]

If true, that's an absolutely horrific mistake to make. Sounds like possibly panicked and shut off the wrong engine not thinking straight.

 

[True__Believer]

If true, that's an absolutely horrific mistake to make. Sounds like possibly panicked and shut off the wrong engine not thinking straight.

We can advance our technology all we like, but as always humans are prone to brain fades. This one cost a lot of lives unfortunately.

Edit: Here is the engine flight data, and it pretty much confirms it:

http://imgur.com/NPa0UAK

Also there were three pilots in the cockpit, with over 28,000 hours flight time combined. Mind blowing.

 

[rdhopkins2]

Just my opinion but you have to question the standard of training. Gut instinct should be to do the right thing (which gets drilled in to you during your original training).

 

[Papa Juggs]

People shouldn't be allowed to fly planes.

First step is taking away people from behind the wheel of cars.

Would you trust a machine to fly you around?

The answer is that the machine can think faster, react faster, take into account all the system status instantaneously, doesn't get prone to fatigue.

The computer might crash though. There's no real difference in that scenario though, if a plane suffers electrical failure you're doomed if there's humans in the cockpit or not.

 

[Aussie Assault]

Wonder if it was a V-22 pilot who forgot he wasn't flying one there?

Not sure if it has been answered further in the thread but wasnt this to do with some heavy cargo not being secured properly causing it to move inside the plane and throwing out the weight distribution upon takeoff?

 

[DrKrieger]

People shouldn't be allowed to fly planes.

First step is taking away people from behind the wheel of cars.

Would you trust a machine to fly you around?

The answer is that the machine can think faster, react faster, take into account all the system status instantaneously, doesn't get prone to fatigue.

The computer might crash though. There's no real difference in that scenario though, if a plane suffers electrical failure you're doomed if there's humans in the cockpit or not.

Cannot disagree more with this.

A machine doesn't think at all, it responds to its sensors based on algorithms. If a sensor goes bad, the machine can't be guaranteed to work properly.

There was an incident years ago where an airliner (might have been a Cathay 747?) lost all three air data computers. So no altitude, no airspeed, no vertical speed. The pilots were able to recognise what had happened. The pilots understood that the values being indicated by all of these instruments (as well as the altitude being transmitted to atc by the transponder) were false values (if a failure of this nature happens, the affected instruments show erroneous values, not a "warning" flag, because the instruments are also machines and respond to the data being input into them).

Most importantly, the pilots then understood that by having power at a certain value, and setting the nose at a certain position (or attitude if you prefer, they could set up a steady steady, safe descent until they broke clear of cloud, at a level they knew was safe because, as humans, they could use the radio to get information on the cloud base in the area from other aircraft/atc.

A computer would have simply crashed, killing a 747 full of people (indeed this would have happened if the pilots didn't disconnect the autopilot).

The gimli glider attempted a landing at an airfield a computer would have either deemed unsuitable, or simply not known about in the first place, and in doing so saved the lives of a 767 full of people.

In the USA, Captain Sully was deemed a hero for landing his aircraft in a river, recognising it as the most suitable place to put the aircraft down. Can a computer do that? Look out the window and pick an area after a complete propulsion failure and determine that it offers the best chance of survival?

Hell, even this ATR accident had 15 people survive, and by the looks of the video and the photos, they survived because the pilots attempted to reach a suitably safe area.

So when you say "if there's an electrical failure you're probably dead anyway" you're not really advocating for a safer system, you're simply substituting one contributing factor to some incidents for another.

If I'm going to die in a plane crash I want to die knowing that up front there is a person/people just as unwilling to die as I am giving it their all to try and save people.

 

[DrKrieger]

Not sure if it has been answered further in the thread but wasnt this to do with some heavy cargo not being secured properly causing it to move inside the plane and throwing out the weight distribution upon takeoff?

Yeah, all the weight moved back, throwing the centre of gravity outside the envelope. Tail no longer has the power to lower to nose and, well

Edit: http://en.m.wikipedia.org/wiki/National_Airlines_Flight_102 is more on the incident in question

 

[Papa Juggs]

In the case of the recent Air Asia crash pilots might have disengaged computers that caused a crash killing 162 DrKrieger

The simple fact is that people would rather have a human kill them than a machine even if humans kill more people more of the time.

To me I only care about cold hard statistics. Let the computer kill me if it's going to kill me 10% less.

 

[rdhopkins2]

Cannot disagree more with this.

A machine doesn't think at all, it responds to its sensors based on algorithms. If a sensor goes bad, the machine can't be guaranteed to work properly.

There was an incident years ago where an airliner (might have been a Cathay 747?) lost all three air data computers. So no altitude, no airspeed, no vertical speed. The pilots were able to recognise what had happened. The pilots understood that the values being indicated by all of these instruments (as well as the altitude being transmitted to atc by the transponder) were false values (if a failure of this nature happens, the affected instruments show erroneous values, not a "warning" flag, because the instruments are also machines and respond to the data being input into them).

Most importantly, the pilots then understood that by having power at a certain value, and setting the nose at a certain position (or attitude if you prefer, they could set up a steady steady, safe descent until they broke clear of cloud, at a level they knew was safe because, as humans, they could use the radio to get information on the cloud base in the area from other aircraft/atc.

A computer would have simply crashed, killing a 747 full of people (indeed this would have happened if the pilots didn't disconnect the autopilot).

The gimli glider attempted a landing at an airfield a computer would have either deemed unsuitable, or simply not known about in the first place, and in doing so saved the lives of a 767 full of people.

In the USA, Captain Sully was deemed a hero for landing his aircraft in a river, recognising it as the most suitable place to put the aircraft down. Can a computer do that? Look out the window and pick an area after a complete propulsion failure and determine that it offers the best chance of survival?

Hell, even this ATR accident had 15 people survive, and by the looks of the video and the photos, they survived because the pilots attempted to reach a suitably safe area.

So when you say "if there's an electrical failure you're probably dead anyway" you're not really advocating for a safer system, you're simply substituting one contributing factor to some incidents for another.

If I'm going to die in a plane crash I want to die knowing that up front there is a person/people just as unwilling to die as I am giving it their all to try and save people.

I totally agree with you. Life is full of managed risks. Pilots, especially in Australia must meet high standards during training and as part of that training are taught manual flying as well as the use of autopilot. The emergency training is just as important as the standard training as pilots must know how to recognise issues and how to solve them safely. At the end of the day, the pilot flying the plane wants to make it back to their loved ones too.

 

[DrKrieger]

I get where you're coming from Juggalo Balla, but I think you would see a lot of events that are solved by pilots today before they have an effect on safety turning into accidents. A computer will never be as good as a person at judging when to attempt a landing and when to hold out and wait as a squall line passes the airport, or hold off on taking off while a flock of birds cross the runway.

Training to a good standard is certainly important, and that's where the judgement of the passenger has to come into play. If the passenger values safety, they'll pay for a ticket with an airline that has a reputation for safety, even if it's more expensive.

My own belief is still that a pilot will safely land an aircraft the majority of time, with a safety issue that would cause a computer controlled aircraft to crash. You're right about wanting a human to crash my plane instead of a machine, but that comes down to my own belief that a human has a better chance of solving the issue than a robot, even if they get it wrong sometimes.

 

[Geelong_Sicko]

Not sure if it has been answered further in the thread but wasnt this to do with some heavy cargo not being secured properly causing it to move inside the plane and throwing out the weight distribution upon takeoff?

Yeah, I was just being smart. A V-22 is a hybrid plane-helicopter and it looked like the transport was trying to make a vertical landing like a chopper (or a Harrier, I guess) would.

 

[rdhopkins2]

Given the quick recovery of the black boxes and aircraft, the final report may take only 6mths or so