Wikiposts
Search
Tech Log The very best in practical technical discussion on the web

Habsheim

Thread Tools
 
Search this Thread
 
Old 18th Feb 2014, 23:44
  #501 (permalink)  
 
Join Date: Jul 2002
Location: UK
Posts: 3,093
Likes: 0
Received 0 Likes on 0 Posts
@CONF iture:

On what evidence are you basing these assertions? To achieve an Alpha Max flypast at or above 100ft RA without Alpha Floor activating and causing the aircraft to climb away, A/THR must be disabled. Therefore there should be no automatic "hunting" of thrust during the manoeuvre, because thrust setting is manual. Any variation of thrust needs to be manually commanded.

As Chris Scott says:
Originally Posted by Chris Scott
The EFCS targets an AoA of alpha-prot with neutral stick, and alpha-max with the stick fully aft. Nz doesn't rule.
The EFCS *targets* Alpha Max with the stick fully aft - it doesn't *guarantee* it.

Where did you see these sorties being performed?

Last edited by DozyWannabe; 18th Feb 2014 at 23:58.
DozyWannabe is offline  
Old 19th Feb 2014, 00:17
  #502 (permalink)  
Thread Starter
 
Join Date: Jan 2008
Location: Blighty (Nth. Downs)
Age: 77
Posts: 2,107
Received 4 Likes on 4 Posts
Quotes from CONF_iture:

"It is not exactly clear why you guys need to introduce notions of "flare maneuver" ..."

Simply because it takes a force to change the trajectory of any vehicle - in this case the FPA. The vertical component of increased thrust alone is not sufficient - extra lift from the wing ** is required. Just for the moment, let's forget about any measures to avoid phugoid oscillations, or whatever. In the case that you are arguing, where the AoA is already at alpha-max, the only way to increase wing lift is to increase the speed. As I have previously reminded you, Valpha-max increases as the load factor (Nz) increases. To put it the other way round, increased speed at alpha-max generates an increase of Nz - which gives an increase of FPA.

In fact, even after the climb has been established, the wing lift requirement will be greater than in level flight, because its vertical component still has to balance the weight. So, at a steady AoA of alpha-max, Valpha-max has to be slightly higher in a climb than in level flight.

"Thrust controls the V/S"

In effect, that's correct, and for the PF to expect the PNF to do that for him would be impracticable.

"Thrust does not control the speed which stays at Valphamax"

For the reasons I have explained, at any given weight, Valpha-max is a variable speed.

** [EDIT]
However, see the reference to "TOTAL aerodynamic lift" in the fourth paragraph of Owain Glyndwr's post (below).

Last edited by Chris Scott; 19th Feb 2014 at 09:34. Reason: ** added.
Chris Scott is offline  
Old 19th Feb 2014, 07:23
  #503 (permalink)  
 
Join Date: Jun 2011
Location: West of Offa's dyke
Age: 88
Posts: 476
Likes: 0
Received 0 Likes on 0 Posts
If I may throw in a few thoughts:

The natural phugoid period is approximately 0.23*Vkts TAS, that is about 28 seconds at 120 kts. If you are looking at a transition period of say 3 seconds that is only going to be one tenth of the phugoid cycle and I doubt it you would see much variation in that time. The terms AI have put into alphaprot to stabilise the trajectory also have a significant short term effect, basically by limiting the allowable AOA if the speed is decreasing but also there may be (probably is) a pitch rate term to add damping to whatever short period motion is associated with alphaprot.

WRT the discussion on what happens transiently if you increase thrust at alphamax I can offer two partial explanations; the reality is probably some combination thereof.

If you add thrust you will, of course, get a nose up pitching moment. In steady state conditions to maintain AOA constant this additional pitch has to be balanced by a nose down pitch from the tail. This means a reduction in tail download so there will be an increase in TOTAL aerodynamic lift even if AOA stays constant. In addition of course you will get an increase in total vertical force from the vertical component of the increased thrust.This is why Vs1g is established with idle thrust.

In the transient the increased thrust will cause the aircraft to pitch upwards and the EFCS will react to this changed pitch rate by applying down elevator (or more accurately will reduce the amount of up elevator) and the mechanism described above will come into play. So the total vertical force will be increased a little by increasing thrust even at constant AOA and airspeed.

The other explanation, which I think would be more powerful, rests on the fact that alphamax is not a 'hard' limit. The generation of wings we are discussing does not have an abrupt 'stall' [I cannot answer for the latest generation with sharklets or curved up winglets/wing tips; they may well have different characteristics]. The likes of the A320 start to have flow breakdown somewhere near midspan and this spreads outboard and inboard as AOA is increased above this point. This is true for both the clean aircraft and with flaps deflected. It is the buffet produced by these separations that becomes the limit to useable lift. In effect alphamax is almost a subjective limit, although formally it may be defined by the level of buffet 'g' at the pilot's station reaching a set level.

Now the aircraft is not going to fall out of the sky, or the simulator come off its mountings if the buffet temporarily goes up a little - assuming that is that the simulator has realistic buffet reproduced anyway. Consequently, there will be a margin of AOA available, but not available for general use, beyond alphamax.

Any closed loop control system will need to preserve some margin between the nominal system maximum and any genuine physical limit. The margin will depend on the nature and consequences of that physical limit. If it is a potentially catastrophic 'cliff edge' limit then the margins will need to be carefully set but if it is, as here, a 'soft' limit then one can be a little more relaxed.

To judge by the Gordon Corps video, AI are quite happy to see AOA go as high as 17 deg with full flaps even though the maximum usable AOA is set at 15 deg.
Owain Glyndwr is offline  
Old 19th Feb 2014, 07:28
  #504 (permalink)  
 
Join Date: Jul 2009
Location: France - mostly
Age: 84
Posts: 1,682
Likes: 0
Received 0 Likes on 0 Posts
Originally Posted by CONF iture
That's how the alpha protection has been designed.
At full back stick the elevators are all for the AoA control to maintain it at alpha max.
That is correct - it is not designed to maintain airspeed. But never mind, OK465 has settled that aspect. Apparently the speed variation is only a few knots for the magnitude of the thrust increase and the rate of change involved.

P.S.
With the natural phugoid period being approximately 28 seconds the speed variation of 'a couple of knots' in that period would be quite slow so it would easily go unnoticed.

Last edited by HazelNuts39; 19th Feb 2014 at 08:56. Reason: P.S.
HazelNuts39 is offline  
Old 19th Feb 2014, 11:51
  #505 (permalink)  
 
Join Date: Jan 2005
Location: W of 30W
Posts: 1,916
Likes: 0
Received 0 Likes on 0 Posts
Originally Posted by Dozy
Any variation of thrust needs to be manually commanded.
And the problem is ... ?

The EFCS *targets* Alpha Max with the stick fully aft - it doesn't *guarantee* it.
A properly flown alpha max demo will deliver alpha max, what make you think the FCS would not comply ?

Originally Posted by Chris Scott
In effect, that's correct, and for the PF to expect the PNF to do that for him would be impracticable.
I disagree on that as well. For the PF to maintain full back stick for a while is not of all comfort. If he wants to concentrate outside, I find it practicable for the PNF to stay inside and manage thrust versus v/s.

Originally Posted by Owain Glyndwr
To judge by the Gordon Corps video, AI are quite happy to see AOA go as high as 17 deg with full flaps even though the maximum usable AOA is set at 15 deg.
That's why it is surprising in the Habsheim case, as alpha max was set at 17.5 deg and thrust reached 83% N1, the elevator had no intention to deliver anything closer than 2.5 deg short of alpha max ...
CONF iture is offline  
Old 19th Feb 2014, 13:14
  #506 (permalink)  
 
Join Date: Jun 2011
Location: West of Offa's dyke
Age: 88
Posts: 476
Likes: 0
Received 0 Likes on 0 Posts
@Confiture

Well of course the elevator doesn't have an intention to deliver anything; it is merely the servant of the control laws. So far as those are concerned the behaviour is not really surprising - you should read the Bilboa report where it says:

Additionally, in these high angle of attack situations, with a tendency towards phugoid movement, studied by longitudinal stability, in which the aeroplane oscillates between two kinetic and potential energy levels, the EFCS behaves as a damper of the oscillations, commanding appropriate variations of angle of attack in a way that, when the aircraft is slowing down, makes it pitch downward and vice versa,"
IIRC in the Habsheim accident, the aircraft was still losing speed up until one second before collision with the trees. Although higher power had been commanded, it was simply set too late to arrest the deceleration and remove the damping correction in time for pilots demand for higher AOA to be satisfied safely and in time.

Last edited by Owain Glyndwr; 20th Feb 2014 at 07:20. Reason: Minor correction
Owain Glyndwr is offline  
Old 19th Feb 2014, 17:24
  #507 (permalink)  
Thread Starter
 
Join Date: Jan 2008
Location: Blighty (Nth. Downs)
Age: 77
Posts: 2,107
Received 4 Likes on 4 Posts
Hi OK465,
Thanks for your detailed, hands-on sim reports. Just to pick up on one:

"The jam engine accels or decels seem to cause the largest AOA variations (both below and above depending) from the stabilized alphamax value, but still resulting in no more than a couple knots above Valphamax on the thrust increases or dipping a couple of knots into the red band on rapid thrust reductions before the available FCS actuated elevator (and AOA control) gets things back under control at Valphamax."

You describe the IAS as varying above or below Valpha-max by "no more than a couple of knots". But did you notice whether the Valpha-max itself remains constant, or if it varies slightly with load factor and/or FPA?

Quote from CONF_iture:
"For the PF to maintain full back stick for a while is not of all comfort. If he wants to concentrate outside, I find it practicable for the PNF to stay inside and manage thrust versus v/s."

So, starting from straight-and-level at alpha-max, let's look at the control of the flight path:
the PF controls bank and can initiate a descent (intentionally or accidentally), but not reverse it;
the PNF controls climb and descent, provided the PF doesn't interfere.

Interesting work-share arrangement, 100ft above the ground. I suppose it might have worked.
Chris Scott is offline  
Old 20th Feb 2014, 10:37
  #508 (permalink)  
Thread Starter
 
Join Date: Jan 2008
Location: Blighty (Nth. Downs)
Age: 77
Posts: 2,107
Received 4 Likes on 4 Posts
Quote from OK465:
"The first obvious change in Valphamax occurs when your wrist gets sore and push forward on the SS resulting in Valphamax noticeably decreasing as the aircraft is unloaded. "

Yes, perhaps you need to alternate between seats! (Being right-handed, I tended to over-control initially on my occasional visits to the R/H seat.) Any further info on variations in Valpha-max (or lack of) would be most helpful.

Returning to the planned Habsheim flypast (at alpha-max, with thrust effectively controlling VS/FPA while the EFCS adjusts pitch to maintain Valpha-max), I've had another look at the degree of control available to the "PF". As you say, if his wrist gets sore and he releases the stick, the a/c pitches down to target the new AoA that the stick has commanded. If the PNF leaves the thrust constant, the IAS will increase to a figure higher than the 1g-Valpha-max. So, as the PF pulls the stick fully back again, there will be a slight excess of IAS available to flare the a/c as the alpha-max is restored. Until the speed bleeds off again, the lift will be higher than before the disturbance, enabling a climb. In other words, it's the old story of potential energy being converted to kinetic, and then back to potential again.

However, the above scenario illustrates the potential for dis-coordination between the two pilots. As Confit says, the PNF's task is to adjust thrust to keep the VS zero. The PF has initiated a descent by allowing the stick to move towards neutral, but the PNF doesn't know that. Therefore, observing the undesired descent, he is likely to increase the thrust just as the PF is flaring the a/c. The thrust increase will have to be reversed immediately if the a/c is to be prevented from overshooting the 100 ft target height because of the increase in total energy.

Had this pilot-duo ever practised the manoeuvre ensemble?
Chris Scott is offline  
Old 20th Feb 2014, 15:11
  #509 (permalink)  
 
Join Date: Jun 2011
Location: Devonshire
Age: 96
Posts: 297
Likes: 0
Received 0 Likes on 0 Posts
In #410 I did suppose ( with no evidence )


" This may have been a display which had been practiced a number of times successfully, perhaps along a standard R/W...
If there were no trees."


To add dual control, with only the PNF able to increase altitude... And at 100ft ( or could we allow 46ft R ?) might be overstressing the co-operation of the two pilots... without their having had a lot of practice.


And their awareness of the trees.
LT
Linktrained is offline  
Old 20th Feb 2014, 19:07
  #510 (permalink)  
 
Join Date: Jan 2005
Location: W of 30W
Posts: 1,916
Likes: 0
Received 0 Likes on 0 Posts
Originally Posted by Owain Glyndwr
you should read the Bilboa report
Interesting you mention Bilbao as the part you did highlight is applicable to pretty turbulent conditions on final approach phase below 200 ft RA when the aircraft encountered strong and changing vertical and horizontal gusts while descending at a rate of around 1,200 ft/min ...

How smooth was Habsheim ...

Originally Posted by Chris Scott
So, starting from straight-and-level at alpha-max, let's look at the control of the flight path:
the PF controls bank and can initiate a descent (intentionally or accidentally), but not reverse it;
the PNF controls climb and descent, provided the PF doesn't interfere.
In Habsheim, as the airplane was never slow enough to get to alpha max for the flypast, the altitude control through thrust management by the PNF did not come to the point it was applicable.
Thrust management by the PNF for vertical speed control during a more conventional alpha max presentation by Airbus is surely a good idea and I'm not too sure what you find 'impracticable' about it ... ?
CONF iture is offline  
Old 20th Feb 2014, 20:01
  #511 (permalink)  
 
Join Date: Jul 2009
Location: France - mostly
Age: 84
Posts: 1,682
Likes: 0
Received 0 Likes on 0 Posts
Phugoid

Just to illustrate what I had in mind when I wrote about a phugoid in post #493.

The following graph shows the variation of speed and height that would result in the following conditions:
- the airplane is initially in level flight at a constant speed of 110 kts TAS
- at t=0 thrust is instantaneously increased to that required for stabilized climb at a flight path angle of 8.5 degrees
- after t=0 thrust and drag are constant and both act in the direction of the flight path, i.e. no thrust component normal to the flight path
- angle of attack is constant, i.e. no damping
- the variation of air density is negligible

This is of course entirely theoretical, but it illustrates the effectivity of the artificial damping in OK465's simulator exercise, combined with the effect of increasing thrust in 2-3 seconds instead of instantaneously and of the thrust component normal to the flight path.

It also illustrates the fundamental difficulty of controlling the flight path with thrust alone, without the ability to control angle of attack.

P.S.
The graph illustrates the point I made in the discussion: the acceleration starts immediately, the flight path changes after some speed increase.


Last edited by HazelNuts39; 21st Feb 2014 at 11:43. Reason: P.S.
HazelNuts39 is offline  
Old 20th Feb 2014, 20:39
  #512 (permalink)  
 
Join Date: Jun 2011
Location: West of Offa's dyke
Age: 88
Posts: 476
Likes: 0
Received 0 Likes on 0 Posts
Originally posted by Confiture

How smooth was Habsheim ...
I am really unsure what point you are trying to make here. AFAIK the same alphaprot law applies no matter what the atmospheric conditions.

Could you please explain what relevance the conditions pertaining to Bilbao have to the behaviour of the system in Habsheim?
Owain Glyndwr is offline  
Old 20th Feb 2014, 21:27
  #513 (permalink)  
 
Join Date: Jan 2005
Location: W of 30W
Posts: 1,916
Likes: 0
Received 0 Likes on 0 Posts
Originally Posted by Owain Glyndwr
AFAIK the same alphaprot law applies no matter what the atmospheric conditions.
Apparently Airbus thought otherwise as following Bilbao they developed a new standard for the ELAC, standard L81, to modify the logic in the AOA protection in case of turbulent conditions.
CONF iture is offline  
Old 21st Feb 2014, 01:32
  #514 (permalink)  
 
Join Date: Jul 2002
Location: UK
Posts: 3,093
Likes: 0
Received 0 Likes on 0 Posts
The post-Bilbao changes related to the phugoid-damping logic, not Alpha Protection. There were no "turbulent conditions" at Habsheim, nor did the EFCS pitch commands at Habsheim reflect what happened at Bilbao.
DozyWannabe is offline  
Old 21st Feb 2014, 05:25
  #515 (permalink)  
 
Join Date: Jun 2011
Location: West of Offa's dyke
Age: 88
Posts: 476
Likes: 0
Received 0 Likes on 0 Posts
Originally posted by Conf_iture

Apparently Airbus thought otherwise as following Bilbao they developed a new standard for the ELAC, standard L81, to modify the logic in the AOA protection in case of turbulent conditions.
True, but if you have read the Bilbao report you will also have been aware that the modification was a deletion of the phase advanced AOA term that was part of the logic that triggered entry into the alphaprot mode [and a change to the logic of alphaprot deselection, but that is not relevant here], not a change to the basic alphaprot laws themselves.

You may also know that this post Bilbao change was actually a reversion to the standard that was applicable at Habsheim.

Air Safety Week July 18 2001
The other aspect governing alpha protection is the rate at which AOA is allowed to change before reaching the protection limit. The alpha protection is triggered by two combined conditions: a threshold AOA and the rate of AOA change. To change the outcome in dynamic wind conditions near the ground, Airbus plans to modify the software to eliminate pitch rate as a controlling factor in alpha protection. In plain language, with the rate of change in the value of AOA being removed, the modification basically reverts the software to an earlier standard where pitch rate was not part of alpha protection (the pitch-rate limitation was installed as a result of post-1988 flight tests).
Despite what Dozy has written I have seen nothing to suggest that the basic phugoid damping terms in alphaprot have ever been changed, (and anyway the phugoid damping is an intrinsic part of alphaprot so his remarks make no sense). Consequently I see no reason why the point I was emphasising:

the EFCS behaves as a damper of the oscillations, commanding appropriate variations of angle of attack in a way that, when the aircraft is slowing down, makes it pitch downward and vice versa
should not be a valid explanation of the reason why alphamax was not developed at Habsheim. I repeat - the necessary thrust increase was applied too late.

Last edited by Owain Glyndwr; 21st Feb 2014 at 06:05.
Owain Glyndwr is offline  
Old 21st Feb 2014, 06:50
  #516 (permalink)  
 
Join Date: Apr 2008
Location: back of beyond
Posts: 116
Likes: 0
Received 0 Likes on 0 Posts
Owain, you and Dozy are more or less saying the same thing in different words.


A rate factor in the feedback loop provides a damping function - think "D" in a PID controller.


Fully agree with your conclusion though, and glad to see some consensus emerging through all the smoke put up about this affair. Great thread people and thanks from a mostly-lurker for all the info and discussion.
fizz57 is offline  
Old 21st Feb 2014, 07:02
  #517 (permalink)  
 
Join Date: Jun 2011
Location: West of Offa's dyke
Age: 88
Posts: 476
Likes: 0
Received 0 Likes on 0 Posts
Fizz,

I didn't think Dozy and I were saying the same thing, because he was suggesting that the phugoid damping logic was changed and I am suggesting that it wasn't.
Nice to know that you agree with the conclusions though.
I think that rate of change of AOA signal was in the forward loop not the feedback. The feedback loop I believe still includes pitch rate as a damper.

Last edited by Owain Glyndwr; 21st Feb 2014 at 07:15.
Owain Glyndwr is offline  
Old 21st Feb 2014, 11:09
  #518 (permalink)  
 
Join Date: Jun 2011
Location: Devonshire
Age: 96
Posts: 297
Likes: 0
Received 0 Likes on 0 Posts
Owain Glyndwr,


" Necessary thrust increase was applied too late..."


Or too low ? Or both ? The planned height was supposed to be 100 ft.

(I do not recall the actual height of the trees.)
Linktrained is offline  
Old 21st Feb 2014, 12:06
  #519 (permalink)  
 
Join Date: Jun 2011
Location: West of Offa's dyke
Age: 88
Posts: 476
Likes: 0
Received 0 Likes on 0 Posts
Linktrained

A fine distinction!

As a non-pilot I would have said he was flying too low and left it too late. so I suppose my answer is "both".

The BEA report gives the average tree height as 12m (39ft)
Owain Glyndwr is offline  
Old 21st Feb 2014, 14:27
  #520 (permalink)  
 
Join Date: Jun 2009
Location: florida
Age: 81
Posts: 1,610
Received 55 Likes on 16 Posts
Interesting graph, Nuts

Although the magnitude of the phugoid seems a bit large, we still see a very rapid increase in FPA and climb rate.

I bet that the pilot would have donated vital parts of his anatomy to gain 100 feet in about 1300 feet of horizontal travel, ya think? 7 seconds at 115 knots +/-

Last edited by gums; 21st Feb 2014 at 14:27. Reason: spelling
gums is offline  


Contact Us - Archive - Advertising - Cookie Policy - Privacy Statement - Terms of Service

Copyright © 2024 MH Sub I, LLC dba Internet Brands. All rights reserved. Use of this site indicates your consent to the Terms of Use.