The Anatomy of a Pitchup!
The full graphs at the end of the report are interesting, thanks for the link. It's difficult without more detail on the "Flight Path Analysis" to draw firm conclusions: it may be that the AOA trace was inferred using some fairly weak assumptions -- it's not from the flight recorder.
I don't think we'll ever know with certainty what happened. The CAB cites the interaction of vertical gusts with control inputs. I find it difficult to see how the vertical gusts could in themselves cause the pitch-up in a way that is consistent with the data. I can certainly see that they could cause confusion and control inputs that made matters worse.
If the aircraft were to encounter a strong downdraft, a pitch-up would be accompanied by a spike of downwards acceleration, until the AOA recovered to its trimmed value. The excursions between 7:00 and 11:00 in the trace show that sort of effect.
You can argue that there's some evidence of that in the traces around 12:05 (though the acceleration never falls below 1g), but you can easily also argue that it was in response to a control input, as has been inferred in the elevator and stick force traces. Between 12:05 and 12:10 there seems to be a nose up control input, and if the aircraft was already above its trimmed airspeed, that's only going to make matters worse.
The pitch-up seems to be a combination of being above trimmed airspeed (whether or not I can persuade you that that was due to a horizontal gust), and a nose-up control input. There's little evidence that it was caused by a downdraft. Regardless of the cause of the pitch-up, what broke the aircraft was the irrecoverable dive induced by the subsequent sustained full nose-down control input.
To address wsherif1's original point, I think the CAB does at least speculate (page 42) on why the nose-down input was held for so long -- a combination of the negative g (perhaps they couldn't even reach the controls?) and the lightening or reversal of stick force at high down-elevator loads on the 720.
I don't think we'll ever know with certainty what happened. The CAB cites the interaction of vertical gusts with control inputs. I find it difficult to see how the vertical gusts could in themselves cause the pitch-up in a way that is consistent with the data. I can certainly see that they could cause confusion and control inputs that made matters worse.
If the aircraft were to encounter a strong downdraft, a pitch-up would be accompanied by a spike of downwards acceleration, until the AOA recovered to its trimmed value. The excursions between 7:00 and 11:00 in the trace show that sort of effect.
You can argue that there's some evidence of that in the traces around 12:05 (though the acceleration never falls below 1g), but you can easily also argue that it was in response to a control input, as has been inferred in the elevator and stick force traces. Between 12:05 and 12:10 there seems to be a nose up control input, and if the aircraft was already above its trimmed airspeed, that's only going to make matters worse.
The pitch-up seems to be a combination of being above trimmed airspeed (whether or not I can persuade you that that was due to a horizontal gust), and a nose-up control input. There's little evidence that it was caused by a downdraft. Regardless of the cause of the pitch-up, what broke the aircraft was the irrecoverable dive induced by the subsequent sustained full nose-down control input.
To address wsherif1's original point, I think the CAB does at least speculate (page 42) on why the nose-down input was held for so long -- a combination of the negative g (perhaps they couldn't even reach the controls?) and the lightening or reversal of stick force at high down-elevator loads on the 720.
Join Date: Aug 2000
Location: UK
Posts: 1,914
Likes: 0
Received 0 Likes
on
0 Posts
Bookworm- I think that is as far in agreement with the events as I can see. Flight recording then was extremely primitive and caution should be attached to these readings.
PickyPerkins, I've been thinking about that excerpt from the CAB report and I think it is simplistic. An updraft will induce an instantaneous weathercocking nose down. Due to the inertia of the aeroplane, it will tend to continue on the same horizontal path with increased AofA. When it comes out of the updraft, its AofA may even be negative, giving a strong pitch up momentum- if it then rapidly moves into downdraft territory, there will be a cumulative effect and it will provide a very powerful increased weathercocking nose up effect. To counter this, anything up to full nose down control may have been applied. I think the catastrophic event occured just after 12.20.
PickyPerkins, I've been thinking about that excerpt from the CAB report and I think it is simplistic. An updraft will induce an instantaneous weathercocking nose down. Due to the inertia of the aeroplane, it will tend to continue on the same horizontal path with increased AofA. When it comes out of the updraft, its AofA may even be negative, giving a strong pitch up momentum- if it then rapidly moves into downdraft territory, there will be a cumulative effect and it will provide a very powerful increased weathercocking nose up effect. To counter this, anything up to full nose down control may have been applied. I think the catastrophic event occured just after 12.20.
Join Date: Jul 2002
Location: 40N, 80W
Posts: 233
Likes: 0
Received 0 Likes
on
0 Posts
I see that when the altitude returns to the same as it was at the beginning of the event (red arrows), the airspeed has also returned to the same or a very slightly lower value.
This means that very little enery was gained/lost overall from the up/down draughts during this period of 32 seconds, although at the end of this time (at 12m 32s) the elevator was full down, the AoA was -11 degrees, the aircraft was pitched 90 degrees down, and the crew was experiencing a negative 2.5 g acceleration. All this with little or no change in total energy.
Isn't almost no change in total energy (potential + kenetic) during altitude excursions a characteristic of a phugoid?
Of course once the crew intervened it ceased to be a simple phugoid, but there was still negligible overall energy input from the up/down draughts.
Cheers,
This means that very little enery was gained/lost overall from the up/down draughts during this period of 32 seconds, although at the end of this time (at 12m 32s) the elevator was full down, the AoA was -11 degrees, the aircraft was pitched 90 degrees down, and the crew was experiencing a negative 2.5 g acceleration. All this with little or no change in total energy.
Isn't almost no change in total energy (potential + kenetic) during altitude excursions a characteristic of a phugoid?
Of course once the crew intervened it ceased to be a simple phugoid, but there was still negligible overall energy input from the up/down draughts.
Cheers,
Last edited by PickyPerkins; 18th Jan 2005 at 03:39.
Join Date: Jul 2002
Location: 40N, 80W
Posts: 233
Likes: 0
Received 0 Likes
on
0 Posts
I thought it might be interesting to try to put together some diagrams of what people on this thread have said ought to happen to an aircraft meeting a short duration down draught.
The diagrams below depict an aircraft flying from left to right, first through still air, then through a down draught, then into still air again.
Case 1
This diagram assumes that the aircraft weathercocks instantaneously to the new relative wind direction as soon as it enters the down draught by rotating about its center of gravity. If there is no change in engine thrust or trim, the lift vector, L, will be greater than in still air because the relative wind speed is greater, and it will act at the angle “a” to the vertical adding to the induced drag.
However I think that the increased lift and drag may be negligible and the aircraft will effectively emerge again into still air at the same altitude and speed that it started with.
For example, the lift, L, in the down draught will be increased over the still air lift by the square of the increase in the relative wind, or by a factor of (v*2+ d*2)/v*2. The vertical component of this lift is [(v*2+ d*2)/v*2][v/((v*2+d*2)*0.5)] = (1+(d/v)*2)*0.5.
We can get an idea of how big this factor (1+(d/v)*2)*0.5 is from some measurements made by NASA in 2000, when they flew an instrumented Boeing 757 through convective turbulence and in two encounters rated as “severe turbulence” measured maximum down draughts of 15.00 m/s and maximum up draughts of 18.41 m/s. However, these flights deliberately avoided direct entry into regions with the strongest radar returns.
If we assume a down draught twice as strong as the strongest measured by NASA, then d=30 m/s and the increased factor derived above (1+(d/v)*2)*0.5 = 1.02 if the aircraft is moving at 280 kts (as was NW 705), i.e. the vertical component of lift increases by only 2% with a down draught twice as strong as the strongest measured by NASA. So as I suggested above, these effects may be negligible and the aircraft may effectively emerge again into still air at the same altitude and speed that it started with.
Case 2
This case assumes that the aircraft takes a finite time to weathercock into the new relative wind.
In the first transition region the plane will experience a jolt downwards which will cause it to start to lose altitude, and during the second transition a very similar jolt upwards which will leave the aircraft flying about horizontally and at about the same speed but at a lower altitude. The wind and lift vectors are essentially the same as in Case 1. The difference between Case 1 and Case 2 is only that the jolt down initiates a loss in altitude which continues until the second, upward, jolt.
Case 3
This case assumes that in addition to everything that happens in Cases 1 and 2, each jolt will start a phugoid, the first initiating a pitch up rotation, while the second initiates a faster, potentially larger amplitude, pitch down rotation (assuming that there is no input from the pilot).
This sequence in Case 3 is reminiscent of the words in the CAB report on the NW 705 incident quoted in an earlier post (but talking about an up draught rather than a down draught):
Up draught causes
an initial weather cocking pitch down into the relative wind
followed “ultimately” by a pitch up and a gain in altitude.
Down draught causes
an initial weather cocking pitch up into the relative wind
followed “ultimately” by a pitch down and a loss in altitude.
I am sure that these diagrams are all too simplistic and that other effects come into play, but these sketches incorporate my impression of what people on this thread have been saying ought to happen when an aircraft flying straight and level meets a short-duration down draught with no input from the pilot.
An extended abstract of the NASA turbulence measurements is available in pdf format at:
http://ams.confex.com/ams/13ac10av/1...acts/40038.htm
Cheers,
The diagrams below depict an aircraft flying from left to right, first through still air, then through a down draught, then into still air again.
Case 1
This diagram assumes that the aircraft weathercocks instantaneously to the new relative wind direction as soon as it enters the down draught by rotating about its center of gravity. If there is no change in engine thrust or trim, the lift vector, L, will be greater than in still air because the relative wind speed is greater, and it will act at the angle “a” to the vertical adding to the induced drag.
However I think that the increased lift and drag may be negligible and the aircraft will effectively emerge again into still air at the same altitude and speed that it started with.
For example, the lift, L, in the down draught will be increased over the still air lift by the square of the increase in the relative wind, or by a factor of (v*2+ d*2)/v*2. The vertical component of this lift is [(v*2+ d*2)/v*2][v/((v*2+d*2)*0.5)] = (1+(d/v)*2)*0.5.
We can get an idea of how big this factor (1+(d/v)*2)*0.5 is from some measurements made by NASA in 2000, when they flew an instrumented Boeing 757 through convective turbulence and in two encounters rated as “severe turbulence” measured maximum down draughts of 15.00 m/s and maximum up draughts of 18.41 m/s. However, these flights deliberately avoided direct entry into regions with the strongest radar returns.
If we assume a down draught twice as strong as the strongest measured by NASA, then d=30 m/s and the increased factor derived above (1+(d/v)*2)*0.5 = 1.02 if the aircraft is moving at 280 kts (as was NW 705), i.e. the vertical component of lift increases by only 2% with a down draught twice as strong as the strongest measured by NASA. So as I suggested above, these effects may be negligible and the aircraft may effectively emerge again into still air at the same altitude and speed that it started with.
Case 2
This case assumes that the aircraft takes a finite time to weathercock into the new relative wind.
In the first transition region the plane will experience a jolt downwards which will cause it to start to lose altitude, and during the second transition a very similar jolt upwards which will leave the aircraft flying about horizontally and at about the same speed but at a lower altitude. The wind and lift vectors are essentially the same as in Case 1. The difference between Case 1 and Case 2 is only that the jolt down initiates a loss in altitude which continues until the second, upward, jolt.
Case 3
This case assumes that in addition to everything that happens in Cases 1 and 2, each jolt will start a phugoid, the first initiating a pitch up rotation, while the second initiates a faster, potentially larger amplitude, pitch down rotation (assuming that there is no input from the pilot).
This sequence in Case 3 is reminiscent of the words in the CAB report on the NW 705 incident quoted in an earlier post (but talking about an up draught rather than a down draught):
Up draught causes
an initial weather cocking pitch down into the relative wind
followed “ultimately” by a pitch up and a gain in altitude.
Down draught causes
an initial weather cocking pitch up into the relative wind
followed “ultimately” by a pitch down and a loss in altitude.
I am sure that these diagrams are all too simplistic and that other effects come into play, but these sketches incorporate my impression of what people on this thread have been saying ought to happen when an aircraft flying straight and level meets a short-duration down draught with no input from the pilot.
An extended abstract of the NASA turbulence measurements is available in pdf format at:
http://ams.confex.com/ams/13ac10av/1...acts/40038.htm
Cheers,
Last edited by PickyPerkins; 20th Jan 2005 at 16:18.
