Sully's Flare on the Hudson: Airbus Phugoid Feedback
Join Date: Apr 2003
Location: Europe
Posts: 3,039
Likes: 0
Received 0 Likes
on
0 Posts
What some people probably don't understand is that alpha-protect is not a normal flying state to be in. It is unthinkable to intentionally enter alpha-protect during an approach to land or ditch. Alpha-protect is a last ditch effort (by the aircraft and no pun intended!) to prevent a stall.
KayPam, Sully did not 'exactly know what he was doing'. He thought he was flying above VLS, i.e. above normal safe approach speed. But in fact he was far, far, very unsafely far below. Read the report. Again, no criticism, just facts.
KayPam, Sully did not 'exactly know what he was doing'. He thought he was flying above VLS, i.e. above normal safe approach speed. But in fact he was far, far, very unsafely far below. Read the report. Again, no criticism, just facts.
Penko, I guess I am with you.
Lack of good accurate speed control suggests to me that the crew were close to overloaded by the circumstances.
Those who suggest that reversion to direct law may have produced a better outcome are really suggesting that, after configuring to direct law the pilot is now capable of flying the aircraft beyond the buffet boundary, at the lift boundary, at the same time avoiding the stall boundary. Something he would have to learn in a one off situation at 150 feet on a power off ditching approach.
I would be somewhat surprised if any pilots, other that Airbus test pilots, would have a clue how close to the stall boundary they are flying, having penetrated the buffet boundary. It is not in the airline pilots job description - nor should it be.
I can make no comment on whether there was sufficient energy remaining for any significant ROD reduction following the direct law scenario. Without knowledge of the characteristics of the CL/Alpha curve it is a big ask to assume any significant amount of extra lift was available by increasing Alpha by a couple of degrees.
Full marks to Sully for making a wise decision to ditch, he was ably supported by Airbus and those who certificated the product, a great result from a most unenviable situation.
Lack of good accurate speed control suggests to me that the crew were close to overloaded by the circumstances.
Those who suggest that reversion to direct law may have produced a better outcome are really suggesting that, after configuring to direct law the pilot is now capable of flying the aircraft beyond the buffet boundary, at the lift boundary, at the same time avoiding the stall boundary. Something he would have to learn in a one off situation at 150 feet on a power off ditching approach.
I would be somewhat surprised if any pilots, other that Airbus test pilots, would have a clue how close to the stall boundary they are flying, having penetrated the buffet boundary. It is not in the airline pilots job description - nor should it be.
I can make no comment on whether there was sufficient energy remaining for any significant ROD reduction following the direct law scenario. Without knowledge of the characteristics of the CL/Alpha curve it is a big ask to assume any significant amount of extra lift was available by increasing Alpha by a couple of degrees.
Full marks to Sully for making a wise decision to ditch, he was ably supported by Airbus and those who certificated the product, a great result from a most unenviable situation.
Last edited by zzuf; 22nd May 2017 at 09:51.
Thread Starter
Join Date: May 2017
Location: San Diego
Posts: 66
Likes: 0
Received 0 Likes
on
0 Posts
No question that Sully was wrong to fly the approach way to slow. The NTSB report and pilots here know that. That said, he had a little pitch-up room i flare which was denied. Next time, or when other landing events trigger alpha-protect as we've seen in the past, is when changes will benefit passengers.
Sully still had 3.5 degrees of pitch before alpha-max, and actual stall doesn't even happen until CL-max at around 5 degrees more than his A320 got. Remember, while close to the ground, with a descent rate, going right to stall in those final 2 seconds would be fine right before impact. Right to the edge, maybe even hitting CL-max and ground effect at the same time, knowing the water was very close. IF Sully would have been able to get 11 or 12 degrees pitch in flare like he was asking for, his descent rate would have been less, and his deck angle would be right at ideal for water impact.
From the Aircraft Performance Study document:
---"Figure 7 shows that between 15:30:36 and the touchdown at 15:30:43, the pitch angle increases from 9.5° to 11° and then settles back to 9.5°, even though in the last two seconds the left longitudinal side stick is at its aft limit, and α is below αmax. "
Straight from the Aircraft Performance Study:
----"A phugoid damping feedback term in the flight control laws, that is active in α−protection mode, attenuated the airplane’s nose-up pitch response to progressively larger aft side stick inputs made below 100 ft radio altitude. "
That damping doesn't belong there so close to the ground in flare. It lowered the nose a bit when he needed a bit of nose-up. Above about 100 ft AGL (RA), its good to have the damping. The damping is there in alpha-protection "general handling" and usually pilots aren't aware of it or almost never use it, which is OK. Its only there in general handling to help out in low airspeed situations.
From the Aircraft Performance Study document:
---"Figure 7 shows that between 15:30:36 and the touchdown at 15:30:43, the pitch angle increases from 9.5° to 11° and then settles back to 9.5°, even though in the last two seconds the left longitudinal side stick is at its aft limit, and α is below αmax. "
----"A phugoid damping feedback term in the flight control laws, that is active in α−protection mode, attenuated the airplane’s nose-up pitch response to progressively larger aft side stick inputs made below 100 ft radio altitude. "
That damping doesn't belong there so close to the ground in flare. It lowered the nose a bit when he needed a bit of nose-up. Above about 100 ft AGL (RA), its good to have the damping. The damping is there in alpha-protection "general handling" and usually pilots aren't aware of it or almost never use it, which is OK. Its only there in general handling to help out in low airspeed situations.
Join Date: Jun 2007
Location: Wanderlust
Posts: 3,403
Likes: 0
Received 0 Likes
on
0 Posts
The aircraft was flown outside the flight envelope because of human limitations under the circumstances and the aircraft behaved as designed. So where is the problem? The speed was dropping all along and without thrust that increases ROD because higher AOA causes more drag. Were it not so the VLS will have no meaning. Trying to pick and choose to connect dots is a waste of time. Another day he may have deactivated the GPWS and would have got speed, speed warning which would have helped him maintain the speed, yet another day he may not have started the APU and reached stall warning at 100ft and pushing the stick forward would have been worse. You try it in SIM in alternate law at 150ft get your speed to stall warning, push the stick and land you will need to stand more than a drink to your buddies.
Last edited by vilas; 22nd May 2017 at 17:20.
Join Date: Feb 2009
Location: Seattle
Posts: 379
Likes: 0
Received 0 Likes
on
0 Posts
What about reduced "ground-speed" for water landing?
There seems to be quite a bit of discussion here about where this airplane was with respect to lower speed limits. I wonder what is known about the relationship between speed across the water and chances of surviving a water landing? When landing on a paved runway (with or without gear down) touchdown sink rate is more important than ground speed. When landing on water, speed across the water might be more important. I would have thought Sully's desire would have been to get as slow as possible short of stalling to reduce the kinetic energy that would be released in short order once on the water. To that extent, Sully might have wanted to be able to command a few more degrees nose up pitch attitude to further reduce speed at touchdown.
Thread Starter
Join Date: May 2017
Location: San Diego
Posts: 66
Likes: 0
Received 0 Likes
on
0 Posts
As a flight control engineer, what do you think about the phugoid damping terms active below 100' AGL radar altimeter? That left Sully with not enough pitch angle, a higher rate of descent, and a higher speed.
A ditching study on the A320 said 11 or 12 theta is fine, a 3-point landing on the water. Of course, too much speed can rip the nacelles off. Too much rate of descent can crack the fuselage.
A ditching study on the A320 said 11 or 12 theta is fine, a 3-point landing on the water. Of course, too much speed can rip the nacelles off. Too much rate of descent can crack the fuselage.
Join Date: Apr 2003
Location: Europe
Posts: 3,039
Likes: 0
Received 0 Likes
on
0 Posts
Read the report, Sully did not want to be as slow as possible, he did not want any of the things you try to ascribe him with.
Sully thought he was above VLS. 'Safely' above VLS. Sully even declined the FO's offer for more flaps (lower stall speed!) arguing it won't make much difference.
Again, I am not criticizing Sully at all, just disputing some of the more miraculous claims on this thread.
(...) the captain stated during postaccident interviews that he thought that he had obtained green dot speed immediately after the bird strike, maintained that speed until the airplane was configured for landing, and, after deploying the flaps, maintained a speed “safely above VLS,” which is the lowest selectable
airspeed providing an appropriate margin to the stall speed. However, FDR data indicated that the airplane was below green dot speed and at VLS or slightly less for most of the descent, and
about 15 to 19 knots below VLS during the last 200 feet.
airspeed providing an appropriate margin to the stall speed. However, FDR data indicated that the airplane was below green dot speed and at VLS or slightly less for most of the descent, and
about 15 to 19 knots below VLS during the last 200 feet.
During postaccident interviews, the captain stated that he used flaps 2 because there were “operational advantages to using flaps 2.” He stated that using flaps 3 would not have lowered the stall speed significantly and would have increased the drag. He stated that he was concerned about having enough energy to successfully flare the airplane and reduce the descent ratesufficiently. He stated that, from his experience, using flaps 2 provides a slightly higher nose attitude and that he felt that, in the accident situation, flaps 2 was the optimum setting.
Join Date: Apr 2003
Location: Europe
Posts: 3,039
Likes: 0
Received 0 Likes
on
0 Posts
Quagmire, again you are putting the cart in front of the horse. Sully did not get those few extra degrees from his Airbus because the Airbus was doing its best not to stall due to Sully's control inputs. You can't first run out of speed, forcing the aircraft in to protect mode and then still expect the aircraft to give you a few extra degrees of pitch for free. A 737 would stall and crash.
The only interesting part in this debate/investigation is that Sully was not aware of the fact that he was in alpha protect. He was not aware how critically low his speed was, probably due to the human factor issues of Ditching in the Hudson. That is something to take away with from this ditching! It could happen to anyone.
The only interesting part in this debate/investigation is that Sully was not aware of the fact that he was in alpha protect. He was not aware how critically low his speed was, probably due to the human factor issues of Ditching in the Hudson. That is something to take away with from this ditching! It could happen to anyone.
Thread Starter
Join Date: May 2017
Location: San Diego
Posts: 66
Likes: 0
Received 0 Likes
on
0 Posts
You keep missing the fact that he still had 3.5 degrees of stall margin remaining at the same time he held full aft stick. And CLmax is at a little higher alpha than AlphaMax.
Nobody disputes that Sully should have had more airspeed. The fact is he didn't, OK, pilot error, we got it. Enough already: Not the topic here. ... The fact is automation saw fit to deny just enough remaining alpha to get the plane to a better deck angle for ditching. That also would mean a lower impact speed could have been obtained. Look into the other several A320 accidents in the past where alpha protection in flare occurred (mentioned in this thread to show the history).
I'd like to find out who thinks its a good idea (like Airbus does) to allow the fly-by-wire to give nose down commands when there is still margin to stall in flare? And what possible reason would there be to limit max lift that way down low in flare?
Again, I do think Airbus's alpha protection is fine above flare altitudes, just not below around 100 feet.
Nobody disputes that Sully should have had more airspeed. The fact is he didn't, OK, pilot error, we got it. Enough already: Not the topic here. ... The fact is automation saw fit to deny just enough remaining alpha to get the plane to a better deck angle for ditching. That also would mean a lower impact speed could have been obtained. Look into the other several A320 accidents in the past where alpha protection in flare occurred (mentioned in this thread to show the history).
I'd like to find out who thinks its a good idea (like Airbus does) to allow the fly-by-wire to give nose down commands when there is still margin to stall in flare? And what possible reason would there be to limit max lift that way down low in flare?
Again, I do think Airbus's alpha protection is fine above flare altitudes, just not below around 100 feet.
Join Date: Dec 2015
Location: France
Posts: 507
Likes: 0
Received 0 Likes
on
0 Posts
Quagmire, again you are putting the cart in front of the horse. Sully did not get those few extra degrees from his Airbus because the Airbus was doing its best not to stall due to Sully's control inputs. You can't first run out of speed, forcing the aircraft in to protect mode and then still expect the aircraft to give you a few extra degrees of pitch for free. A 737 would stall and crash.
The only interesting part in this debate/investigation is that Sully was not aware of the fact that he was in alpha protect. He was not aware how critically low his speed was, probably due to the human factor issues of Ditching in the Hudson. That is something to take away with from this ditching! It could happen to anyone.
The only interesting part in this debate/investigation is that Sully was not aware of the fact that he was in alpha protect. He was not aware how critically low his speed was, probably due to the human factor issues of Ditching in the Hudson. That is something to take away with from this ditching! It could happen to anyone.
alpha max is 17.5 in conf 2.
The max angle of attack that was reached was 15°.
Alpha max in itself includes a margin as well.
So there was more than enough margin.
Sully was probably aware of how low his speed was, thanks to the appropriate information on the PFD (speed trip with orange/red colouring), but maybe he actually wanted to have a forward speed as low as possible.
Join Date: Sep 2016
Location: USA
Posts: 803
Likes: 0
Received 0 Likes
on
0 Posts
He was not aware how low his speed was, and he did not want what he had. (He wanted what he erroneously thought he had)
"the captain stated
during postaccident interviews that he thought that he had obtained green dot speed immediately
after the bird strike, maintained that speed until the airplane was configured for landing, and,
after deploying the flaps, maintained a speed “safely above VLS,” which is the lowest selectable
airspeed providing an appropriate margin to the stall speed. However, FDR data indicated that
the airplane was below green dot speed and at VLS or slightly less for most of the descent, and
about 15 to 19 knots below VLS during the last 200 feet."
Thread Starter
Join Date: May 2017
Location: San Diego
Posts: 66
Likes: 0
Received 0 Likes
on
0 Posts
Wait a second. I just found something buried on page 194 of the final NTSB Accident Report appendix. Mini-scandal. Yellow alert. Call the Lone Gunmen of the X-Files.
Starting on Page 194 (appendix area) of the NTSB Report, it shows the BEA (Bureau d’Enquetes et d’Analyses’) requested the NTSB reverse their initial indictment of the phugoid damping in flare, saving Airbus from criticism. And the NTSB complied. Not if I was there. Must have upset a least a couple of NTSB pilots or flight control engineers on the team.
Originally, the early NTSB reports mentioned it all. Page 194 of the NTSB Report (the BEA French section) says:
"However, this accident demonstrates that, by offsetting the pilot's ANU sidestick inputs, the phugoid-damping feedback function of the alpha-protection mode could make flaring the airplane to attain the recommended ditching touchdown parameters more difficult." --- That statement didn't make it into the final report, but was in the appendix as a reported document change.
Then the BEA makes a strange remark: "In reality, phugoid oscillations induce pitch variations that can have more severe consequences than a high vertical speed when entering the water."
Except for the fact that the flight is over way, way, before any phugoid oscillation can develop. Its flare for pete's sake. The BEA made a valid case for higher AGL use of the damping terms, but not for close to the water or ground during descent at flare.
Starting on Page 194 (appendix area) of the NTSB Report, it shows the BEA (Bureau d’Enquetes et d’Analyses’) requested the NTSB reverse their initial indictment of the phugoid damping in flare, saving Airbus from criticism. And the NTSB complied. Not if I was there. Must have upset a least a couple of NTSB pilots or flight control engineers on the team.
Originally, the early NTSB reports mentioned it all. Page 194 of the NTSB Report (the BEA French section) says:
"However, this accident demonstrates that, by offsetting the pilot's ANU sidestick inputs, the phugoid-damping feedback function of the alpha-protection mode could make flaring the airplane to attain the recommended ditching touchdown parameters more difficult." --- That statement didn't make it into the final report, but was in the appendix as a reported document change.
Then the BEA makes a strange remark: "In reality, phugoid oscillations induce pitch variations that can have more severe consequences than a high vertical speed when entering the water."
Except for the fact that the flight is over way, way, before any phugoid oscillation can develop. Its flare for pete's sake. The BEA made a valid case for higher AGL use of the damping terms, but not for close to the water or ground during descent at flare.
Only half a speed-brake
The captain declared he was well above the speed. The data trace indicate the speed was significantly below any meaningful target. (more than 15 knots, that is a LONG distance on the speedtape)
KayPam, how do you condense the two statements above into "was probably aware"???
KayPam, how do you condense the two statements above into "was probably aware"???
All this chat about alpha margin is totally irrelevant without knowledge of the extra manoeuvre capability to be gained by an alpha increase. We all know that the speed was way below a safe speed and I would guess that all the swept wing jet pilots here would know that any similar aircraft would have extremely limited manoeuvre capability at this speed.
Without a total CL/Alpha curve for the aircraft on the day (and total drag curve) most of this discussion is speculation. The only thing that really matters to the pilot is how much extra G he can pull without making the situation even worse.
With reference to the phugoid damping:
1. In alpha protect Airbus found that that phugoid could be excited, at an alpha=alpha protect (phugoid is essentially a constant alpha oscillation).
2. For whatever reason, Airbus decided this characteristic was unacceptable and included a damping term in the flight control laws.
3. While, generally speaking, the phugoid is a non event because there wlll be adequate longitudinal control authority to stop the oscillation during either the pitch up or pitch down phase.
4. However, if you put in a hard controllability limit (like alpha protect) the pilot will be unable to stop the oscillation during the pitch down phase - he is denied the ability to increase alpha by the control laws.
5. Best to allow the flight control protection laws (but not the pilot) a little extra authority in terms of alpha capabilty so this oscillation can be prevented.
6. It has been suggested to give the pilot a "little more authority" to be able to demand a slightly higher alpha by removing the phugoid damping term. Seem to me this just moves the phugoid to a higher alpha and in no way guarantees that the pilot will be able to stop the phugoid if it triggers in the nose down mode, at the new alpha protect, when on final approach to his ditching.
7. Seems to me that if you think alpha protect is a good idea, but you don't wish to damp the phugoid you must accept that the nose down pitch phase will not be controllable.
8. If you give the pilot sufficient authority to stop the phugoid at any time, by definition you no longer have alpha protect.
Without a total CL/Alpha curve for the aircraft on the day (and total drag curve) most of this discussion is speculation. The only thing that really matters to the pilot is how much extra G he can pull without making the situation even worse.
With reference to the phugoid damping:
1. In alpha protect Airbus found that that phugoid could be excited, at an alpha=alpha protect (phugoid is essentially a constant alpha oscillation).
2. For whatever reason, Airbus decided this characteristic was unacceptable and included a damping term in the flight control laws.
3. While, generally speaking, the phugoid is a non event because there wlll be adequate longitudinal control authority to stop the oscillation during either the pitch up or pitch down phase.
4. However, if you put in a hard controllability limit (like alpha protect) the pilot will be unable to stop the oscillation during the pitch down phase - he is denied the ability to increase alpha by the control laws.
5. Best to allow the flight control protection laws (but not the pilot) a little extra authority in terms of alpha capabilty so this oscillation can be prevented.
6. It has been suggested to give the pilot a "little more authority" to be able to demand a slightly higher alpha by removing the phugoid damping term. Seem to me this just moves the phugoid to a higher alpha and in no way guarantees that the pilot will be able to stop the phugoid if it triggers in the nose down mode, at the new alpha protect, when on final approach to his ditching.
7. Seems to me that if you think alpha protect is a good idea, but you don't wish to damp the phugoid you must accept that the nose down pitch phase will not be controllable.
8. If you give the pilot sufficient authority to stop the phugoid at any time, by definition you no longer have alpha protect.
Last edited by zzuf; 23rd May 2017 at 01:40.
Thread Starter
Join Date: May 2017
Location: San Diego
Posts: 66
Likes: 0
Received 0 Likes
on
0 Posts
QuagmireAirlines
I would be interested in how you would intend to wash out the phugoid damping term, from the alpha protect law, at low altitude, over a very short time interval, while there is a good chance that the side stick is being held fully aft, and not create even worse handling problems. Seems like it could be a fun day in the engineering sim.
You don't seem to have addressed if, at the speed available, there is any advantage in terms of manoeuvre in even attempting to increase alpha.
Your stylized CL/Alpha curve may show the relationship of various alpha limits, but in no way does it represent the A320 CL/Alpha curve, nor does show just how much usable CL was left available.
I would be interested in how you would intend to wash out the phugoid damping term, from the alpha protect law, at low altitude, over a very short time interval, while there is a good chance that the side stick is being held fully aft, and not create even worse handling problems. Seems like it could be a fun day in the engineering sim.
You don't seem to have addressed if, at the speed available, there is any advantage in terms of manoeuvre in even attempting to increase alpha.
Your stylized CL/Alpha curve may show the relationship of various alpha limits, but in no way does it represent the A320 CL/Alpha curve, nor does show just how much usable CL was left available.
Quote Airbus response to NTSB
There are feedbacks within the AoA protection law aiming at damping the phugoid mode (low frequency mode). Without these feedbacks, an aircraft upset from its stabilized flight point up to constant high AoA would enter a phugoid (which is, by definition, a constant AoA oscillation) without possibility to stabilize the trajectory. As a consequence, commanded AoA is modulated: for instance, if aircraft speed is decreasing and/or pitch attitude is increasing, pilot's commanded AoA is lowered in order to avoid such a situation to degrade.
Trying to run simulation without such damping features on the very last seconds of the flight, without considering what could have been the effect such features brought upstream during the flight on the overall Aircraft trajectory and management by the crew would be pure speculation, as not supported by technical facts.
On the last 10 sec in the air of Flight 1549 , DFDR data show that pitch attitude is increasing and CAS decreasing. Then, the phugoid damping terms are non null and are acting in the sense to decrease the finally commanded AoA vs. the stick command, in order to prevent the Aircraft from increasing the phugoid features.
......
..... with a loss of engine thrust, as in Flight 1549, the aircraft energy management significantly increases the pilot workload. Under these circumstances, aircraft is still able to reach the optimum water impact configuration, but this is a demanding task which requires time and significant pilot focus. Typically, the flare initiation height will be critical to the achievement of the optimum water entry conditions.
Would appear that Airbus have no enthusiasm for the "turn off" phugoid damping just prior to the flare scenario.
There are feedbacks within the AoA protection law aiming at damping the phugoid mode (low frequency mode). Without these feedbacks, an aircraft upset from its stabilized flight point up to constant high AoA would enter a phugoid (which is, by definition, a constant AoA oscillation) without possibility to stabilize the trajectory. As a consequence, commanded AoA is modulated: for instance, if aircraft speed is decreasing and/or pitch attitude is increasing, pilot's commanded AoA is lowered in order to avoid such a situation to degrade.
Trying to run simulation without such damping features on the very last seconds of the flight, without considering what could have been the effect such features brought upstream during the flight on the overall Aircraft trajectory and management by the crew would be pure speculation, as not supported by technical facts.
On the last 10 sec in the air of Flight 1549 , DFDR data show that pitch attitude is increasing and CAS decreasing. Then, the phugoid damping terms are non null and are acting in the sense to decrease the finally commanded AoA vs. the stick command, in order to prevent the Aircraft from increasing the phugoid features.
......
..... with a loss of engine thrust, as in Flight 1549, the aircraft energy management significantly increases the pilot workload. Under these circumstances, aircraft is still able to reach the optimum water impact configuration, but this is a demanding task which requires time and significant pilot focus. Typically, the flare initiation height will be critical to the achievement of the optimum water entry conditions.
Would appear that Airbus have no enthusiasm for the "turn off" phugoid damping just prior to the flare scenario.
Last edited by zzuf; 23rd May 2017 at 08:47.
Join Date: Jun 2007
Location: Wanderlust
Posts: 3,403
Likes: 0
Received 0 Likes
on
0 Posts
Phugoid basically are pitch oscillations at high AOA. Manually trying to control that when they are rapid usually will lead to being out phase and aggravating it. That is why they are done automatically something like yaw damping. Pilot has no business to enter that regime willingly. In 447 the aircraft was flown into flight envelope regime never experienced even by a test pilot and normal pilots and non pilots kept discussing/are still discussing modifications to recover from that regime without any data assigning blame to the design. AF447, Hudson is done and dusted no one in authority is going take any notice about the same issues over and over again.
