Is Pilot Error to Blame for Icing Condition Accidents?

[Ronnin]

I heard a pilot saying that too much flap in this situation could lead to a tailstall. I thought extended flaps anytime you need more lift.

 

[edward]

The plane was on auto pilot until just before the crash. Auto pilot has a feature commonly called "stick shaker stick pusher"

As a plane approaches stall speed the Yoke will vibrate then within seconds if no action is take by the pilot the yoke will push forward to increase speed.

There is now some speculation that the pilot may have been alarmed by the forward motion of the yoke and pulled back too far resultiing in the initial unexplained 30 degress + upward nose pitch.

 

[Danger]

It seems to me that everyone is overlooking something. You're all concentrating on icing affecting the wings and lift. Let's not forget that it takes even less ice to screw up the horizontal stabilizer, which controls attitude. It's possible that the pilot had no pitch control and didn't realize it until George kicked out and handed it back to him.
If the ice had also jammed the elevators, he might have horsed back on the yoke and broken it loose, which might have resulted in a higher nose position than intended.

 

[russ_watters]

I wonder why the air speed would have been too low?

Did the airspeed decrease before or after he pulled back on the stick? With the aerodynamics destroyed, there is a lot of room to speculate on what the plane would do, but the ice could have contributed to the violence of the pullup and if the pullup caused an immediate stall (a stall is caused by a too-high angle of attack, not a low speed), the airspeed would drop rapidly and the plane would go from flying to falling like a brick in just a few seconds. I'm interested to see how much detail we get about what the plane was doing those few seconds - it may be quite revealing.

And once the airspeed gets pretty low, there is no possibility for control left. The violent rolling and pitching might not be what you would envision a flying airplane to do - it could be more like a leaf or a piece of paper fluttering to the ground.

 

[mgb_phys]

With the aerodynamics destroyed, there is a lot of room to speculate on what the plane would do,

The stick-shake stall warnign is presumably based on the design stall speed of the plane, not the stall speed of the new ice+aerofoil wing.
Is there any way for the system sense how close you are to stall from the real time response of the plane to controls?

 

[edward]

The stick-shake stall warnign is presumably based on the design stall speed of the plane, not the stall speed of the new ice+aerofoil wing.
Is there any way for the system sense how close you are to stall from the real time response of the plane to controls?

I don't think so that is why they recommend that the pilot be in control during icing conditions. The auto pilot compensates for conditions a pilot would have already felt in the controls.

There is also an issue with the landing gear being lowered was it the pilot or the autopilot?

Apparently it was the pilot who raised the gear again a few seconds later.

 

[Cyrus]

WOW!

45 degree turn WHILE a 45 degree nose down IS INSANE!

When we practice 45 degree turns we call it 'steep turns' and it is STEEEP. YOU FEEL THE G's in that kind of a turn. They were CRANKING that airplane LIKE HELL. They could have very easily just ripped the god damn wings off!

Danger: Thats a GREAT analysis I didn't think of! I emailed it to my professor at NASA langely. I'll post his reply when he sends it. He does crash analysis as well on the side. I'll also ask my helicopter professor who also does crash stuff.

 

[WhoWee]

What is the difference between a "wing stall" and a "tail stall" and how should you correct each...I heard an aviation expert discussing it on the radio and I didn't catch all of the details...I thought he said the correction of one is the opposite of the other?

 

[FredGarvin]

http://aircrafticing.grc.nasa.gov/courses/inflight_icing/related/3_2_3f_RI.html

The gist with a tail stall is related to flap usage because the wings are still producing lift with the decreased speed but the tail can not. It seems like a design flaw to have a portion of the flight envelope that can have the wings still producing lift and the tail section not. The icing in this case could have created that in that the icing screws up the aero of the tail so it stalls sooner than it normally would. The cool things about the video above are:
1) The statement that even though you correct very quickly, the aerodynamics has a lag and takes some time. Hence the statement that if you have this happen on approach, you're toast.
2) The flight video shows a stalled stabilizer in flight.
3) You can see how violent the yoke can become in a real stall.

MGB...to answer your question, NO they are calibrated for a clean wing.

Here's a nice quote that shows, I think, that the pilot had the right idea, but the wrong implementation:

William Rieke, a NASA research pilot and chief of aircraft operations at NASA Glenn, says to consider the 180-degree turn only if you have no other options -- like an airport in front of you. "If you are getting ice and you have to do a 180 to get out of it and in the process go back through [the ice] and accumulate more -- or worst case, if the air mass is moving in the direction you just turned to, you may be in it even longer than your initial encounter." If terrain clearance or clouds were not an issue, Rieke would prefer to descend 3,000 feet or, if the airplane was powerful enough to climb with the weight and drag of accumulated ice, climb 3,000 feet in an effort to get out of the icing conditions. (Most training aircraft would not be able to climb if they were carrying an appreciable quantity of ice.)

They were at a low altitude within 7 miles of the airport. Undoubtedly the guy was lining up for final. He was thinking just get this thing down ASAP. Unfortuantely ASAP in this case was too soon. This will be supported when they release the configuration of the airplane and it's speed just prior to the crash.

http://flighttraining.aopa.org/ft_magazine/fullstory.cfm?id=5222&issue_title=February

 

[edward]

Radical steep turns are sometimes used to dislodge the ice from wings. Even the B-1 bomber pilots were taught to use the method. They usually crashed. The B-1, built to fly at supersonic speed and at tree top level over Siberia had no deicing system.

 

[rcgldr]

http://aircrafticing.grc.nasa.gov/courses/inflight_icing/related/3_2_3f_RI.html

The gist with a tail stall is related to flap usage because the wings are still producing lift with the decreased speed but the tail can not.

Listen to the chat in the actual test part of the video. The recovery procedure was to retract the flaps which resulted in near instant recovery. The issue wasn't speed (not with a pitch down attitude more than -10 degrees), but that the deployed flaps were interfering with the tail control at that air speed for that particular aircraft. This could be due to the air flow at the tail being interfered with, or because increasing camber by deploying flaps increases the pitch down torque on the wings.

update - The issue was speed related in the sense that the elevator could not provide enough sufficient negative lift (upwards force) to prevent the aircraft from pitching downwards, even at maximum non-stall deflection. The situation is made worse by the pilot increasing the elevators angle of attack so that it stalls (due to angle of attack, not air speed). In the video, the flaps are deployed while the aircraft has a pitch of -13 degrees. The pilot continously pulls back on the elevator to reduce the amount of negative pitch, which decreases the air speed. The flaps produced enough drag that even with pitch around -9 degrees, the airspeed continues to decrease as the pilot feeds in more up elevator until the elevators angle of attack becomes excessive and stalls (which is independent of air speed), severely reducing the negative lift at the tail and allowing the aircraft to quickly pitch downwards to -30 degrees. Retracting the flaps reduces the camber related pitch down torque on the wings, and allows the aircraft to pick up speed more quickly, restoring control authority to the elevator. The downwards diversion of air from the wings with fully extended flaps could also be reducing the air speed at the tail. This is why some aircraft use high "T" tails to keep the control surfaces out of the downwash from the wing.

For a normal aircraft, the center of gravity is front of the center of pressure, and the tail generates less lift or negative lift in order to maintain pitch angle. It's a form of self-stability in that a given elevator angle will correspond to a given air speed. For a given elevator angle, too much speed pitches aircraft up, slowing it down, too little speed pitches it down, speeding it up, and the aircraft will tend to fly at a particular air speed for a given elevator angle (assuming power output is not changed).

A tail stall means the tail can't generate sufficient negative lift, so the air craft pitches down as shown in the video above. This wouldn't explain the roll reaction described in the actual accident.

Although icing could have been an issue, I've wondered if the accident could have been due to faulty flap and/or air brake deployment, either malfuction (one flap deploying more than the other, or both deploying too much plus air brakes), or pilot error. Another possibility is that the de-icing system failed, and if the pilot tried to reduce air speed to normal flight mode, the aircraft would have stalled (wing stall).

I got the impression that things didn't go bad until the flaps were deployed. As far as auto-pilot usage goes, in zero visibility conditions, the military will use auto-pilot to do the approach and landings (called zero-zero), and some commercial airliners have the same auto-plilot landing capability.

http://en.wikipedia.org/wiki/Autopilot

 

[FredGarvin]

Listen to the chat in the actual test part of the video. The recovery procedure was to retract the flaps which resulted in near instant recovery. The issue wasn't speed (not with a pitch down attitude more than -10 degrees), but that the deployed flaps were interfering with the tail control at that air speed for that particular aircraft. This could be due to the air flow at the tail being interfered with, or because increasing camber by deploying flaps increases the pitch down torque on the wings.

Umm...the issue is somewhat about speed. All aircraft with flaps have Vne's with the flaps deployed in each position which is usually quite a bit slower than most cruising speeds. The fact that the flaps are deployed means that you are operating at reduced speeds.

Did you listen to the interview at the beginning? It was said that the corrective action was applied in about .2 seconds and it too "a few seconds" to recover. I counted about 3 seconds between pitch down and recovery and that's with someone with their hands on the flap controls. That is not instantaneous. They lost 300 feet in altitude and they knew exactly when they were going to induce the stall. Now put someone in the situation where they don't know it's coming and add a much delayed reaction time plus, probably, a lot worse condition of icing.

 

[Cyrus]

http://video.google.com/videoplay?docid=2238323060735779946&ei=HCSeSZaFIKGG_AG94qjsBg&q=nasa+icing&hl=en

Gooooooooooooooooooood video.

 

[Cyrus]

Umm...the issue is somewhat about speed. All aircraft with flaps have Vne's with the flaps deployed in each position which is usually quite a bit slower than most cruising speeds. The fact that the flaps are deployed means that you are operating at reduced speeds.

Did you listen to the interview at the beginning? It was said that the corrective action was applied in about .2 seconds and it too "a few seconds" to recover. I counted about 3 seconds between pitch down and recovery and that's with someone with their hands on the flap controls. That is not instantaneous. They lost 300 feet in altitude and they knew exactly when they were going to induce the stall. Now put someone in the situation where they don't know it's coming and add a much delayed reaction time plus, probably, a lot worse condition of icing.

Based on the NASA video, Jeff is exactly right.

 

[Danger]

What is the difference between a "wing stall" and a "tail stall" and how should you correct each... ...I thought he said the correction of one is the opposite of the other?

The one main point of confusion to groundhogs is that the tailplane produces a negative lift to counteract the aeroplane's natural tendency to pitch down. Most people assume that it acts in the same lift fashion as a main wing. In a wing stall, you want to drop the nose and/or add power in order to recover because your angle of attack is too high. In a tail stall, you have to pull up to regain proper attitude because the angle is too low.

 

[rcgldr]

Listen to the chat in the actual test part of the video. The recovery procedure was to retract the flaps which resulted in near instant recovery. The issue wasn't speed (not with a pitch down attitude more than -10 degrees), but that the deployed flaps were interfering with the tail control at that air speed for that particular aircraft. This could be due to the air flow at the tail being interfered with, or because increasing camber by deploying flaps increases the pitch down torque on the wings.

update - The issue was speed related in the sense that the elevator could not provide enough sufficient negative lift (upwards force) to prevent the aircraft from pitching downwards, even at maximum non-stall deflection. The situation is made worse by the pilot increasing the elevators angle of attack so that it stalls (due to angle of attack, not air speed). In the video, the flaps are deployed while the aircraft has a pitch of -13 degrees. The pilot continously pulls back on the elevator to reduce the amount of negative pitch, which decreases the air speed. The flaps produced enough drag that even with pitch around -9 degrees, the airspeed continues to decrease as the pilot feeds in more up elevator until the elevators angle of attack becomes excessive and stalls (which is independent of air speed), severely reducing the negative lift at the tail and allowing the aircraft to quickly pitch downwards to -30 degrees. Retracting the flaps reduces the camber related pitch down torque on the wings, and allows the aircraft to pick up speed more quickly, restoring control authority to the elevator. The downwards diversion of air from the wings with fully extended flaps could also be reducing the air speed at the tail. This is why some aircraft use high "T" tails to keep the control surfaces out of the downwash from the wing.

The main point of my previous post is that a stall is the result of excessive angle of attack, not airspeed.

http://en.wikipedia.org/wiki/Stall_(flight)

In a wing stall, you want to drop the nose and/or add power in order to recover because your angle of attack is too high. In a tail stall, you have to pull up to regain proper attitude because the angle is too low.

Update - you're correct, I thought "tail stall" mean a true stall of the elevator, but it doesn't. See my post below.

I've corrected this post to read "true stall of elevator".

In both cases, the angle of attack is too high. "True stall of elevator" only occurs when the elevator is excessively deflected upwards (or downwards if inverted flight, I'm ignoring inverted flight for the rest of this) quite a bit. Pulling up further will just worsen the situation. The recovery procedure is to feed in down elevator and/or increase air speed, reducing the angle of attack.

An extreme example of wing stall is a snap roll induced through excessive up elevator at speed without any aileron input. I flown some radio control models that do this, you peg the elevator stick back and instead of climbing, the model just rolls, a true snap roll (one wing stalls a bit before the other, resulting in a very fast roll). A not so obvious case occurs when ground towing a glider. The glider will snap roll if excessive up elevator is used. The first response to any roll reaction during a ground tow, both real and model gliders, is down elevator and then aileron to correct since there isn't enough time to figure out if the issue is a normal roll or snap roll. It's more of an issue for model gliders since there are contests (F3J) where minimum time during launch is a goal, so very high towing forces are used (over 20 g's). Snap rolls can also unintentionlly result from pylon racing models, when the high g turns cause the main wing to stall.

 

[Danger]

Jeff, I have the utmost respect for you as a scientist and educator, but I'm talking about real aircraft as opposed to toys.
A tail stall refers to the inability of the horizontal stabilizer to provide proper trim. That results in a nose-down pitch attitude. Pushing forward on the stick, as you would for a wing stall, will send you straight into the ground.

 

[Cyrus]

update - The issue was speed related in the sense that the elevator could not provide enough sufficient negative lift (upwards force) to prevent the aircraft from pitching downwards, even at maximum non-stall deflection. The situation is made worse by the pilot increasing the elevators angle of attack so that it stalls (due to angle of attack, not air speed). In the video, the flaps are deployed while the aircraft has a pitch of -13 degrees. The pilot continously pulls back on the elevator to reduce the amount of negative pitch, which decreases the air speed. The flaps produced enough drag that even with pitch around -9 degrees, the airspeed continues to decrease as the pilot feeds in more up elevator until the elevators angle of attack becomes excessive and stalls (which is independent of air speed), severely reducing the negative lift at the tail and allowing the aircraft to quickly pitch downwards to -30 degrees. Retracting the flaps reduces the camber related pitch down torque on the wings, and allows the aircraft to pick up speed more quickly, restoring control authority to the elevator. The downwards diversion of air from the wings with fully extended flaps could also be reducing the air speed at the tail. This is why some aircraft use high "T" tails to keep the control surfaces out of the downwash from the wing.

The main point of my previous post is that a stall is the result of excessive angle of attack, not airspeed.

http://en.wikipedia.org/wiki/Stall_(flight)

In both cases, the angle of attack is too high. Tail stalls only occur when the elevator is excessively deflected upwards (or downwards if inverted flight, I'm ignoring inverted flight for the rest of this) quite a bit. Pulling up further will just worsen the situation. The recovery procedure is to feed in down elevator and/or increase air speed, reducing the angle of attack.

An extreme example of wing stall is a snap roll induced through excessive up elevator at speed without any aileron input. I flown some radio control models that do this, you peg the elevator stick back and instead of climbing, the model just rolls, a true snap roll (one wing stalls a bit before the other, resulting in a very fast roll). A not so obvious case occurs when ground towing a glider. The glider will snap roll if excessive up elevator is used. The first response to any roll reaction during a ground tow, both real and model gliders, is down elevator and then aileron to correct since there isn't enough time to figure out if the issue is a normal roll or snap roll. It's more of an issue for model gliders since there are contests (F3J) where minimum time during launch is a goal, so very high towing forces are used (over 20 g's). Snap rolls can also unintentionlly result from pylon racing models, when the high g turns cause the main wing to stall.

No, you are wrong. Danger is right. Watch the NASA video I linked.

 

[rcgldr]

A tail stall refers to the inability of the horizontal stabilizer to provide proper trim. That results in a nose-down pitch attitude.

I wasn't aware that the term "tail stall", isn't the same as a true elevator stall. After doing a web search, a "tail stall" refers to the case where external forces deflect the elevator downwards, yanking the yoke forwards, generally coinciding with deployment of flaps at high speed (probably related to the pitching torque I mentioned before). The pilot has to pull back on the yoke with a lot of force, to overcome the external forces on the tail, to recover the elevator back to a normal position. Although this is not a true "stall" situation (excessive angle of attack on the elevator) it's called a "tail stall".

In the NASA video, it appears that a true elevator stall is induced. Full flaps are deployed at high speed and the pilot states that he applies up to 80 pounds of pulling force on the yoke (he's pulling on a force sensor connected to the yoke) to oppose the external downwards force on the elevator, allowing the pilot to continue to control the aircraft while in a "tail stall" situation. In this case, the pilot continues deflecting the elevator further upwards as the aircraft slows, apparently acheiving a true stall at the elevator, resulting in the aircraft suddenly pitching down to -30 degrees while the yoke is held in place by the pilot (as opposed to being yanked forward). Note that the pilot retracted the flaps as soon as the true stall was detected in order to recover.

I don't know if icing was an issue in the NASA video, but ice accumulation on the horizontal stabilizer and elevator can result in a true stall of the elevator occurring at a lower angle of deflection than normally required to cause a stall on a "clean" tail section.

Full flap deployment at speed is an issue in the design of aircraft. During flight testing of a certain GA aircraft intended for certification, the first time that Norm Howell dropped the flaps, the horizontal tail stalled. (I don't know if this was a true stall or just insufficient negative lift due to inadequate elevator). Norm says that the airplane instantly went over on its back and started an inverted spin. Cool Norm reached over, retracted the flaps and then recovered from the spin. Eventually, in the wind tunnel we found that the airplane needed a cambered (upside down) airfoil on the horizontal tail.:

http://yarchive.net/air/tail_stall.html

Recovering from a tail stall: Here is how to recover from a tail stall: Immediately raise flaps to the previous setting. Pull aft on the yoke. Copilot assistance may be required. Reduce power if altitude permits; otherwise maintain power. Do not increase airspeed unless it is necessary to avoid a wing stall.:

http://www.aopa.org/asf/publications/sa11.pdf

 

[Danger]

I wasn't aware that the term "tail stall", isn't the same as a true elevator stall.

That would certainly explain the original disagreement. My apologies if I seemed condescending.

 

[rcgldr]

I wasn't aware that the term "tail stall", isn't the same as a true elevator stall.

That would certainly explain the original disagreement. My apologies if I seemed condescending.

No problem, my fault for not looking up tail stall before posting. I never heard the term "tail stall" before, I have read a lot about aerodynamics (I also fly radio control gliders). I was aware of how icing can badly affect airfoil shapes, but was totally unaware that deploying flaps at speed, combined with icing effects, could result in tail stall. After looking up more web pages about this, some aircraft are subceptible to tail stall while others aren't (those others still have icing issues, but not the pitch down issue related to tail stall).

Anyway, it doesn't appear that "tail stall" was an issue. The pilot might have mistakenly though the aircraft was in a "tail stall" due to the auto pilot pitching the nose down and shutting off, and pulled back on the yoke in what is reported at a 2G manuever. The news is reporting that the aircraft pitched up 31 degrees (not down) and hitting a 2 g peak (1 g of this is gravity related). It's also possible that the pilot was trying to avoid a mid-air collision with what might have appeared to be another aircraft.

Aircraft pitched up 31 degrees, rolled 106 degrees, tail icing (tail stall) not an issue for this aircraft, but it was on the previous type of aircraft that that pilot had flown a possible reason that the pilot pitched the nose up so much. There's a animated mulitmedia link in the second photo on this web page describing what happened (airspeeds aren't known yet and I don't know why):

http://www.nytimes.com/2009/02/19/nyregion/19crash.html?hp

This web page includes more info, including some air speeds:

http://buffalopundit.wnymedia.net/blogs/archives/8070

It's really too soon to tell. Pitching the nose up 31 degrees from a nose down condition and pulling 2g's seems extreme, causing me to wonder if some type of mechanical failure was involved.

 

[Nick M]

Properly equipped aircraft pass through icing conditions all the time. The key phrase here being "pass through", and not sustained flight. Usually this happens when you're climbing through a layer of icing conditions to clear conditions at a higher altitude, or when descending through an icing layer to land. Even small personal aircraft equipped for passing through icing conditions have de-icing equipment like heated props, inflatable wing boots, and a heat-plate for the windshield so you have a small rectangle for visual reference of the ground. You can't make gear like de-icing boots automatic, because they only work when activated with ice build-up (you can actually make things worse by cycling them without a sufficient ice layer).

I would also question why the pilot didn't notice a change in the flight characteristics of the aircraft if it was indeed icing related. I have my Private Pilot license and a few hours of instrument training (read beginner), but even I can feel the difference between an aircraft at gross weight operating at a high density altitude, and one with just me in it with half-tanks on a cool morning.

Regional airlines kinda scare me anyways. Usually their ranks are filled with greenhorns working insane hours under intense pressure.

Of course it's still safer to fly than to get in your car and turn onto the highway.

 

[mgb_phys]

You can't make gear like de-icing boots automatic, because they only work when activated with ice build-up (you can actually make things worse by cycling them without a sufficient ice layer).

They only work for a certain rate of ice build up, they don't handle ice forming on the tail. Some pilots turn the wing boots on and off in the belief that it's more effective

I would also question why the pilot didn't notice a change in the flight characteristics of the aircraft if it was indeed icing related.

It appears that it was on autopilot. As with most of the low altitude turboprop+ice crashes - the pilot didn't know anything until the plane became unflyable

 

[rcgldr]

deicing bootsThey only work for a certain rate of ice build up, they don't handle ice forming on the tail.

Looking at the diagrams, the leading edges of the vertical and horizontal stablizer also have de-icing boots. The news reports claims that tail icing isn't an issue for this particular aircraft.

pilot not aware of problemIt appears that it was on autopilot. As with most of the low altitude turboprop+ice crashes - the pilot didn't know anything until the plane became unflyable

That isn't known. The aircraft pitched up to 31 degrees, generating 2g's of lift during the process so the wings were able to generate a lot of lift and the elevator generate sufficient downforce to pitch the noseup. What isn't known is why the aircraft pitched up to 31 degrees.

 

[edward]

pilot not aware of problemIt appears that it was on autopilot. As with most of the low altitude turboprop+ice crashes - the pilot didn't know anything until the plane became unflyable

That isn't known. The aircraft pitched up to 31 degrees, generating 2g's of lift during the process so the wings were able to generate a lot of lift and the elevator generate sufficient downforce to pitch the noseup. What isn't known is why the aircraft pitched up to 31 degrees.[/QUOTE]

The plane was on autopilot according to the FTSB. The 31% pitch up was a result of the pilots action. At this point we can only assume that the pilot reacted to the stick pusher by applying full power and pulling back on the yoke.

The stick pusher engages before a plane is actually in a stall situation.

 

[mgb_phys]

The stick pusher engages before a plane is actually in a stall situation.

It engages before the plane reaches it's design stall speed, if the ice changed the aerodynamics so that the stall speed of the new wing shape was higher it could have stalled before there was any warning.