Habsheim

[CONF iture]

Originally Posted by HN39

That is a strange twist - many posts ago I first suggested that phugoid damping was a possible explanation for the 'attenuated' response of the airplane to full back sidestick.

... and I did not oppose it, but what you confirm here is that the phugoid damping (if it is actually the guy behind the restriction ...) was useless and detrimental performance wise.

Bechet did much better without it than Asseline with it.

[Owain Glyndwr]

Looking at the totality of the evidence from the four graphics presented in the BEA report it seems to me that the situation is as follows:

The HAP mode of the EFCS has no knowledge of what the engines are doing. Indeed, with thrust control switched to manual no part of the EFCS uses thrust information. Consequently any aircraft motions produced by thrust changes are treated in exactly the same way as perturbations coming from any other “outside” source such as atmospheric disturbances.

The HAP mode contains logic that in effect says: “Do not allow pitch to increase if the speed is falling”. Introduced to ensure trajectory stability in the phugoid mode, it has its own logic – if flying at or near alphamax with a falling airspeed the last thing one wants to do is allow changes that might take the aircraft over that limit and nearer to alphastall.

The “job” the EFCS was being asked to do has been stated as “Provide the commanded alphamax”, but although at first reading that seems a straightforward objective, it is in fact simplistic. A more accurate task description would be “Provide the commanded alphamax in a controlled manner so as to minimise the risk and potential magnitude of overshoots above alphamax”

The details vary a little depending on the timing of thrust and sidestick applications, but basically when starting from a condition of level flight in the HAP operative AOA range with idle thrust and a reducing airspeed and applying maximum thrust and full back stick the sequence of events would be as follows:

Because the speed is initially reducing, the HAP logic requires that a down elevator signal be added to cancel the pilot’s nose up command. At this point the thrust is low and increasing only slowly, so the pitch remains constant whilst the speed builds up slowly.
As soon as the thrust has increased enough to arrest the deceleration, the HAP logic changes to allow the aircraft to pitch up. By this time though,the thrust has increased enough to provide a noticeable nose up pitching moment so the pilot’s demand can be satisfied without much change to the elevator command. At the end of this phase the speed is still building up slowly and the aircraft has just started to climb. [Note**]
The engines are by now accelerating quite rapidly, being on the steep part of the “S” curve, and if left uncorrected the associated pitching moments would increase the pitch rate to a level where there would be a very real risk of overshooting alphamax. The HAP therefore applies corrective elevator (actually a reduction of ‘up’ elevator) as shown on the graphs.
The aircraft, at this point, is pitching up towards alphamax and starting to climb more steeply.
As the engines near TOGA power their acceleration tails off along with any increase in pitching moment. The HAP elevator command depends on the details of the pitch rate and the proximity to alphamax at this point. If it seems that the pitch rate could be safely increased then the HAP might apply a small nose up correction, but if the pitch rate means that the aircraft is approaching alphamax too rapidly it might need to do the reverse to avoid an overshoot.
It is quite difficult to see exactly what is driving the elevator at this time because it is trying to control the difference between pitch and flight path angle, but we don’t have direct information on either alpha or FPA from the simulator records, although it could in principle be calculated from the data available.

Under this line of reasoning the only substantive difference between any of the four “flights” would be that F-JFKC’s was truncated (literally!)

[Note**] It was at about this point in the sequence that F- GFKC entered the trees.

[DozyWannabe]

Originally Posted by CONF iture

but what you confirm here is that the phugoid damping (if it is actually the guy behind the restriction ...) was useless and detrimental performance wise.

Not quite - as OG says:

Originally Posted by Owain Glyndwr

The “job” the EFCS was being asked to do has been stated as “Provide the commanded alphamax”, but ... [that] is in fact simplistic. A more accurate task description would be “Provide the commanded alphamax in a controlled manner so as to minimise the risk and potential magnitude of overshoots above alphamax”

This is because it's an airliner, not a fighter - HAP mode is designed to maintain stability and trajectory, and Alpha Max is the upper AoA limit of that mode - it is *not* necessarily for use as an escape manoeuvre. In fact it can only really be expected to perform that function in concert with either sufficient airspeed, sufficient manual thrust applied (and available) or with Alpha Floor enabled. Asseline set the situation up so as to have neither.

Bechet did much better without it than Asseline with it.

As the graphs demonstrate (and OG has helped describe), Bechet initially got to a higher AoA in less time, but the overall performance was more or less similar. Whichever way you slice it, the fundamental aspect that cannot be ignored is that, possibly as a result of task saturation, Asseline did not recognise the danger in time and therefore did not act swiftly and decisively enough. That in itself is not the main reason he shoulders the responsibility he does - it's more to do with his decisions prior to then, when he had ample opportunity to turn around and do the run with more margin for error - but instead elected to press on and continue improvising with an ever-narrowing safety margin.

Airbus's own Alpha Max demonstrations were performed with sufficient thrust to achieve and maintain Alpha Max while preserving trajectory, and they were properly prepared such that they had a clear path ahead of the aircraft while in that state, because in that mode and at higher alphas, the room for evasive manoeuvre is significantly reduced.

[CONF iture]

Originally Posted by Owain Glyndwr

The HAP mode contains logic that in effect says: “Do not allow pitch to increase if the speed is falling”.

Bechet got no restriction to obtain more than 17 deg of pitch even if he was under high AoA protection 1 sec after selecting TOGA.
True that speed was bizarrely not falling ... especially compared to Asseline who kept his sidestick in the same position for 3 more seconds ...

if flying at or near alphamax with a falling airspeed the last thing one wants to do is allow changes that might take the aircraft over that limit and nearer to alphastall.

If flying at or near alpha max, there is no such thing as "falling airspeed", but there is flying at or near Valphamax.
"Falling airspeed" would be associated to rapid increase of alpha.

If it seems that the pitch rate could be safely increased then the HAP might apply a small nose up correction, but if the pitch rate means that the aircraft is approaching alphamax too rapidly it might need to do the reverse to avoid an overshoot.

It would do that but brief temporary overshoot of alpha max is part of the protection too as demonstrated by the Airbus test pilot.

Originally Posted by Dozy

This is because it's an airliner, not a fighter - HAP mode is designed to maintain stability and trajectory, and Alpha Max is the upper AoA limit of that mode - it is *not* necessarily for use as an escape manoeuvre.

To the contrary it is specifically designed with that in mind.

In fact it can only really be expected to perform that function in concert with either sufficient airspeed

Alpha max is necessarily associated to "sufficient airspeed" called Valphamax.

As the graphs demonstrate (and OG has helped describe), Bechet initially got to a higher AoA in less time, but the overall performance was more or less similar.

Anything between 5 and 10ft is a lot of performance ... if it's all it takes to survive the bush ...

[DozyWannabe]

Originally Posted by CONF iture

True that speed was bizarrely not falling ...

Are you sure? Annexe X - page 9, right?

If so, then I'm seeing from "Remise des gaz"/full back-stick:

• An immediate increase in pitch for approximately 1 second (about the time it takes to transition between pitch command and HAP mode)
• The speed continues to decay from approx. 120kts to 113kts
• As the mode transition finishes (at approx. 115kts), the pitch increase stops, then actually decreases slightly as the speed decays
• As the speed builds up from the lowest value of 113kts (presumably due to the thrust kicking in), the pitch angle starts to increase again

To the contrary it is specifically designed with that in mind.

Only in conjunction with sufficient airspeed and/or thrust. Alpha Floor is intended to provide a degree of fail-safe functionality in the event that airspeed/thrust is insufficient, which is why turning it off or otherwise preventing its function is not advisable unless precautions regarding minimum altitude and a clear path ahead of the aircraft are strictly adhered to.

Alpha max is necessarily associated to "sufficient airspeed" called Valphamax.

Yes, but unless you know how that's calculated, it's something of a moot point.

Anything between 5 and 10ft is a lot of performance ... if it's all it takes to survive the bush ...

Page 9 indicates that Bechet's altitude was in fact slightly lower than Asseline's at the point of impact, but even if you transpose it to match it doesn't really suggest that it would have made much difference.

[EDIT : Actually, hold that thought - I just remembered something from a while back (cheers for this info Chris)...

Originally Posted by Chris Scott

The four rad-alt TRx antennae on the A320 were mounted on the bottom of the rear fuselage, just forward of the fuselage tapering area ... At the final pitch angle of about +15, the tail-cone and tapering part of the rear fuselage would have been the first to touch any horizontal series of treetops.

The RA of Bechet's pass diverges significantly below that of Asseline's roughly around the same time Bechet gets his initial "bump" in pitch (before HAP mode/damping arrests it as a result of an airspeed delta of approx. -5kts/sec). I'm wondering if this increase in divergence is down to the higher pitch angle and consequent lower physical altitude of the RA antennae. ]

Also, as others have pointed out, the increased pitch angle would have meant that the empennage made contact with the trees lower down and at a more acute angle, increasing the risk of the airframe breaking up before it came to rest and a significant increase in casualties as a result.

[CONF iture]

Yes I am sure that speed was bizarrely not falling ... especially compared to Asseline who kept his sidestick in the same position for 3 more seconds ...
And obviously you don't have any explanation to justify that so different attitudes obtained by Bechet and Asseline are supposedly still producing an identical speed decrease ... ?

Originally Posted by Dozy

Alpha Floor is intended to provide a degree of fail-safe functionality in the event that airspeed/thrust is insufficient, which is why turning it off ...

Asseline applied TOGA before alpha floor would have kicked in ... what's you point again ?

Page 9 indicates that Bechet's altitude was in fact slightly lower than Asseline's at the point of impact, but even if you transpose it to match it doesn't really suggest that it would have made much difference.

Knowing that Bechet started the procedure 10ft lower than Asseline did, it does suggest that it could have made all the difference.

Also, as others have pointed out, the increased pitch angle would have meant that the empennage made contact with the trees lower down and at a more acute angle, increasing the risk of the airframe breaking up before it came to rest and a significant increase in casualties as a result.

Just another typical unsubstantiated catastrophic scenario à la Airbus.

[HazelNuts39]

Originally Posted by CONF iture

True that speed was bizarrely not falling ... especially compared to Asseline who kept his sidestick in the same position for 3 more seconds ...

On the timescale of the chart on page 9 of BEA's 'Additif' (synchronized at the time the flight path passes through 100 ft RA ?), Bechet advanced the thrustlevers between 1 and 1.5 seconds earlier than Asseline.

"Falling airspeed" would be associated to rapid increase of alpha.

Actually, at nominally 1g, with IAS falling at 1.7 kt/second, AoA is increasing at 0.5 degrees/second.

Anything between 5 and 10ft is a lot of performance ... if it's all it takes to survive the bush ...

Knowing that Bechet started the procedure 10ft lower than Asseline did, it does suggest that it could have made all the difference.

The accident report puts the impact at about 5 seconds after the thrust levers were advanced. Bechet's radio altitude was 27 ft when he advanced the thrust levers, and 5 seconds later it was again 27 ft. Asseline's radio altitude was 32 ft when he advanced the thrust levers.

[roulishollandais]

Finaly, which was the math model ?

[CONF iture]

Originally Posted by HN39

On the timescale of the chart on page 9 of BEA's 'Additif' (synchronized at the time the flight path passes through 100 ft RA ?), Bechet advanced the thrustlevers between 1 and 1.5 seconds earlier than Asseline.

I thought the graph was supposed to be a "Recoupement from Habsheim" but are you now suggesting there is in fact not much to be 'recouped' ... ?

Bechet's radio altitude was 27 ft when he advanced the thrust levers, and 5 seconds later it was again 27 ft.

I thought you had earlier established the gain at 10 ft ... ?

Actually, at nominally 1g, with IAS falling at 1.7 kt/second, AoA is increasing at 0.5 degrees/second.

And if you increase your AoA by 5 degrees in 2 seconds what would be the IAS falling rate ... ?

[DozyWannabe]

@CONF - I notice you're avoiding the question over whether the slight increase in pitch would account for the marginally lower RA value given that the increase in pitch angle would cause the rear-mounted RA antennae to be physically closer to the ground.

The IAS drops from 120 at the point Bechet set TOGA to 113kts 5 seconds later. How is that "not falling", let alone "bizarrely"?

And regarding "unsubstantiated catastrophic scenario[s]", it's no such thing - it's basic physics. A flat-ish sheet of metal subjected to an opposing physical force will bear a greater degree of that force as the angle becomes more acute.

@roulis - I'm mystified as to what you actually mean there.

[HazelNuts39]

Originally Posted by CONF iture

I thought the graph was supposed to be a "Recoupement from Habsheim" but are you now suggesting there is in fact not much to be 'recouped' ... ?

The report discusses the difficulty of reproducing simultaneously all elements of the accident flight. My understanding of "Recoupement" is that the objective of Bechet's simulator exercise was to 'replay' the accident flight. The point you are criticising does not concern Bechet's simulation as such, but is about how Asseline's trajectory has been plotted on the graph for comparison with the simulated flight.

I thought you had earlier established the gain at 10 ft ... ?

Actually I said 'less than 10 ft' here. I do not recall exactly how I arrived at that number, but I may have meant to say that Bechet gained 'less than 10 ft' on Asseline in those 5 seconds. Apologies if my wording was imprecise.

And if you increase your AoA by 5 degrees in 2 seconds what would be the IAS falling rate ... ?

If you increase your AoA the loadfactor increases. The IAS continues to fall until the thrust has increased sufficiently to accelerate the airplane.

[roulishollandais]

Originally Posted by DozyWanabee

@roulis - I'm mystified as to what you actually mean there./

mystified by me is surely to much. And what about in my posts ? Math model of june 1988 A320 ? Metalanguage ?
Aren't you from the World best Country in logic ? Charles Babbagge, Ada Lovelace, Lewis Carroll, Alan Matison Turing ! France is it for math !

[CONF iture]

Originally Posted by HN39

The point you are criticising does not concern Bechet's simulation as such, but is about how Asseline's trajectory has been plotted on the graph for comparison with the simulated flight.

I have actually a LOT to criticize :

1. All the graphs are of the lowest standard. I would not have wasted any time on them if you didn't mention the one on page 9. I do think the BEA made its best effort to make sure nobody would be tempted to read them.
2. Those flight and simulator tests were purely smoke screens - what was needed was only the DFDR data to be compared with the most detailed programming of the FCS.

If you increase your AoA the loadfactor increases.

Load factor increase should reveal a change of path, from level flight to ... climb.

The IAS continues to fall until the thrust has increased sufficiently to accelerate the airplane.

I still disagree with your concept.
Kinetic energy is transformed in potential energy until the thrust has increased sufficiently to confirm the early but short term positive climb as IAS now remains stable at Valphamax.

[HazelNuts39]

Originally Posted by CONF iture

Kinetic energy is transformed in potential energy

Indeed, as shown below.

Kinetic energy is transformed in potential energy until the thrust has increased sufficiently to confirm the early but short term positive climb as IAS now remains stable at Valphamax.

No. The airplane will start a phugoid motion at alphamax as shown earlier, unless the pilot relaxes the sidestick and maintains the appropriate pitch attitude.

In the following graph KEht and TEht are the kinetic and total energies, respectively, expressed as a height. Total energy is the sum of kinetic and potential energy.

[DozyWannabe]

@roulis - I've explained at least half a dozen times how the software development work was done, most recently in post #740.

The chances of there being a logical error due to a mis-key - as you theorise - are practically zero because the code is not keyed in manually. The chances of there being a logical error aren't much greater because of the pioneering work on testing, regression and reliability metrics. After that there's the fact that every major system "block" was implemented twice using syntactically different methods, and each is cross-checked against the other.

Even CONF iture seems to be accepting the possibility that it might just be a known feature of the system at this point, even if he doesn't much like how that particular mode functioned.

Between OG and HN39's sterling work, we're looking at a very reasonable explanation of how the aircraft functions would have behaved in that scenario.

[Owain Glyndwr]

@roulis

I have to say that I share DW's puzzlement when trying to understand your point. From previous postings:

A simulator stays a simulator today, since they always have commands which don't exist on the plane and use metalanguage

In the simulator a hidden mistake ...... may modify the behaviour of the simulated system (plane, inertial system, aso) without you know it, and you have a bad surprise when you do the same action on the plane, despite your aerodynamic algorithm is OK.

Finally, which was the math model ?

You seem to be saying that you believe there could have been a mistake in programming the simulator [Dozy has been commenting however on errors in the flight EFCS units]


However, I cannot relate to the simulator having "commands" which don't exist on the aeroplane. The simulators at Toulouse are linked to the "Iron Bird" which uses actual flight standard hardware (and its associated software) so the processing of pilot commands is identical on the aircraft and in the simulator. The only difference is that the simulator uses a mathematical model of the aerodynamics and processes the pilot commands in conjunction with that model to produce the aircraft motions which are fed back into the flight standard EFCS units rather than those emanating directly from aircraft response.To quote from the Airbus FAST article on the iron bird:

From the flight deck, the Iron Bird can be flown like a standard aircraft, with a computer generating the aerodynamic model and such environmental conditions as air density, air temperature, airspeed and Mach number.

Since you seem to be accepting that the likelyhood of error in the aerodynamic representation is small, the only source of difference between simulator and actual aircraft that fits your argument is, so far as I can see, the solution of the six degree of freedom equations of motion. Such an error would surely not be confined to the particular case we have been discussing, and in any case the correspondence between the actual flight records and the simulated version of the last few seconds of flight (page 13 of Annexe X of the BEA report) says that no significant error existed.
Also of course the method of solution of these equations has been been around a long time in various simulators so there is no innovation there.


Could you perhaps be a little more specific regarding the reasons why you think the simulator results to be suspect?

[CONF iture]

Originally Posted by HN39

No. The airplane will start a phugoid motion at alphamax as shown earlier, unless the pilot relaxes the sidestick and maintains the appropriate pitch attitude.

So according to Bechet the airplane has no issue going to alpha max but then, what do you make of the Airbus statement that "the flight protection system operated perfectly and prevented the aircraft from crashing nose first" ?

Also not too sure how you get Valphamax at 113kt ... ?

[HazelNuts39]

So according to Bechet the airplane has no issue going to alpha max

Sorry, I should perhaps have written "phugoid motion at approximately alphamax". The phugoid damping terms will reduce the AoA slightly below alphamax when the airplane is decelerating, and increase AoA when the airplane is accelerating, as conceptually illustrated in the chart below.

... but then, what do you make of the Airbus statement that "the flight protection system operated perfectly and prevented the aircraft from crashing nose first" ?

Well, the flight protection system certainly 'prevented the aircraft from crashing nose first'. As to 'operating perfectly' the BEA report notes:

1.16.1.4.2. Réponse de l'avion à une sollicitation à la profondeur à très basse vitesse

La difficulté à reproduire une manœuvre provient du fait qu'il faut se placer dans des conditions identiques ou très voisines de lois de pilotage impliquant le respect de tous les paramètres influents (facteur de charge ou assiette, incidence, gradient de vitesse, positions de manche en profondeur) alors que ces paramètres ne sont enregistré qu'à une cadence faible sur le DFDR (un point par seconde en général) et qu'une restitution exacte nécessiterait de l'ordre de 40 points par seconde.

In short, the recorded parameters are too few and too coarse to permit a reconstruction of the operation of the flight control system in a complex situation with variations of windspeed and angle of attack encountered by the airplane as it flew below treetop height in the wake of the forest.

Also not too sure how you get Valphamax at 113kt ... ?

I stand corrected: Valphamax should have been the Vs1g of 110 kt.

[DozyWannabe]

Originally Posted by CONF iture

So according to Bechet the airplane has no issue going to alpha max but then, what do you make of the Airbus statement that "the flight protection system operated perfectly and prevented the aircraft from crashing nose first" ?

In addition to HN39's remarks, I should add that "operating perfectly" in this context seems to refer to the fact that it operated as designed - and based on the analysis both of the report and the work done on this thread, that looks like a reasonably accurate statement.

What Bechet seems to have proved (and others here have explained how and why), is that the aircraft will attain Alpha Max as long as speed is sufficient, but the time taken to do so depends on the flight conditions and the mode. Specifically, the time taken to achieve it will be longer if the command (i.e. full back-stick) is given with the aircraft already in HAP mode and decelerating than it will if the aircraft is in pitch command mode when the command is given.

From a systems perspective this makes sense as, when presented with a decelerating aircraft, HAP mode has to constrain rate of movement in the pitch axis (and elsewhere) in order to fulfil the part of its design brief that demands flightpath stability and trajectory be paramount. Pitch command mode is obviously free of those constraints because airspeed deltas are immaterial if pitch (as opposed to AoA) is what's being commanded by the stick.

As to what impact the earlier increase in AoA might have made in that situation - well, we can only speculate. But my speculation based on reading of the graphs (aided immeasurably by other posters' explanations) definitely indicates that there would still have been some impact with the trees, and at that point there are just too many variables involved (particularly with regard to what part of the aircraft hit the trees, at what angle and what damage that would cause) to make even an educated guess. That said, based on the photographic evidence, I do think that the actual path of the aircraft caused the main gear and later the main spar to absorb more of the impact force from the trees, lessening the subsequent forces on the more fragile empennage and tailcone.

[CONF iture]

Originally Posted by HN39

I stand corrected: Valphamax should have been the Vs1g of 110 kt.

But again, why do you still persist equalling Valphamax and stalling speed ... ?
And if you do think Valphamax is Vs1g at 110 then what is Vstall ?

True that all these speeds, all of them, including Vls Valphaprot, should have been published in the report for the weight config altitude at the time ... Where are they ?

For the graph, things don't add up.
How do they or you get identical speed traces from so different attitudes ?
An obvious penalty on the speed there must be by pulling the nose up ... but also a benefit to the altitude ... Where are the ?