Autothrust behaviour with corrupted airspeeds

[Hyperveloce]

Hi there
I am in search of the basic principles of the autothrust and its different modes/settings.
If you are in high altitude cruise, approaching a tropical storm cell with possible turbulences (but not using the "severe turbulences" procedure putting the A/THR off and fixing N1), what would be the selected mode ? Speed mode ?
In speed mode, with airspeeds fluctuating down to a largely understimated value (caused by a ram air port freezing of the Pitots for instance), would it be possible that the A/THR augment the thrust just before disengagement ? (what would be the time constant of the thrust control loop ?). If yes, is it possible for the crew not to notice this thrust level increase ? even in max thrust ?
Are there known exemples of unusual behaviours of the autothrust systems both on Airbuses or Boeings ?
Jeff

[shortfuel]

I am in search of the basic principles of the autothrust and its different modes/settings.

You can try this:
AIRBUS 320 / FCOM 3 / CHAPTER 22 AUTOFLIGHT (http://www.smartcockpit.com/pdf/plane/airbus/A320/systems/0025/)

If you are in high altitude cruise, approaching a tropical storm cell with possible turbulences (but not using the "severe turbulences" procedure putting the A/THR off and fixing N1), what would be the selected mode ? Speed mode ?

If you have no other choice than approaching that cell...you might elect to decelerate towards your best penetration speed due to mod/sev turb expectations. On the bus, in most cases, you will be in Mach mode.

[Jimmy Do Little]

If you are in high altitude cruise, approaching a tropical storm cell with possible turbulences (but not using the "severe turbulences" procedure putting the A/THR off and fixing N1), what would be the selected mode ? Speed mode ?

Probably "Mach." Assuming the "High Altitude Cruise" part of your question and that speed is Managed, per the Airbus SOP (However unlikely that the company SOP may be different in this regard).

In speed mode, with airspeeds fluctuating down to a largely underestimated value (caused by a ram air port freezing of the Pitots for instance), would it be possible that the A/THR augment the thrust just before disengagement ?

I'm not entirely clear about what you're asking. But, with a downward speed trend, the A/Thr will react promptly (Somewhat) in "SPEED MODE" (Mach Mode with Soft Altitude engaged will be a slightly slower reaction), to the point of A/prot protection engagement (Airbus). Next, would be a pitch down command and TOGA commanded.


If yes, is it possible for the crew not to notice this level increase ? even in max thrust ?

Possible, yes. Likely, no.

Are there known examples of unusual behaviors of the auto-thrust systems both on Airbuses or Boeing's ?

I think a BA 777 crew could answer that question!

you might elect to decelerate toward your best penetration speed due to mod/sev turb expectations..


Just an FYI.....

Airbus FCTM 4.010

"...If moderate turbulence is encountered, the flight crew should set the AP and A/THR to ON with managed speed.

If severe turbulence is encountered, the flight crew should keep the AP engaged (Note: Good luck with that as the "G" load would probably keep knocking it out). Thrust levers should be set to turbulence N1 (Refer to QRH), and the A/THR should then be disconnected. (Interesting note: No reference to selected speed [obvious reason] in either FCTM or QRH)..."

Apparently, Airbus trains for two turbulence scenarios - Up to Severe turb (Everything Managed), and Severe or greater (Manage everything, but manual thrust).

Cheers

[Hyperveloce]

Thank you both for these answers.
In the AF 447 case, the A/THR was engaged till the long sequence of problems began at 02:10Z (by its disengagement, with the AP, etc...).
The procedure "severe turbulence" was not applied (A/THR off, fixing N1).
Since the problems were already ongoing at 02:10Z when the A/THR disengaged (an air data module had been probably voted out leading to an ADR disagreement a few minutes later), the airspeeds have probably rolled back just before the A/THR disengagement and I try to know whether the A/THR could have compensated this airspeed drop (by augmenting the thrust/a gentle nose down) just before it was disengaged. I would need to have an idea of the time constant of the A/THR control loop (the control loop seems to be a PID).
If the A/THR did augment the thrust before being disengaged at 02:10Z, it probably did it just before, in the few seconds before, so the crew might have been distracted from this by the long sequence of ECAM fault reports ? Well... before speculating about this possibility, I should try to confirm/infirm that the A/THR augmenting the engine thrust due to corrupted airspeeds is possible in the conditions of interest (AF 447 penetrating a MSC around ~02:00Z).
I found an occurrence of an A/THR being involved into an incident with an A340 in Cayenne but in approach (not in cruise):
"Le vent de face atteint un pic à 17 h 45 min 22 puis diminue très rapidement (11) ; la vitesse air décroît. A 17 h 45 min 26, l’auto-poussée commande une augmentation de poussée mais la vitesse air continue de décroître du fait de la diminution continue du vent et du temps d’accélération des réacteurs. La poussée atteint 40 % une seconde et demie après et 58 % trois secondes après le début d’accélération."
the surface winds were blowing strongly when they suddendly calmed, as seen from the plane, the airspeeds decreased and the A/THR reacted to this by augmenting the thrust to 40% in 1.5 sec., 58% in 3 sec. (the lagging is due to the engines, not to the A/THR: it seems able to respond quickly).

[shortfuel]

I would need to have an idea of the time constant of the A/THR control loop (the control loop seems to be a PID).

I doubt you find that info in any pilot training doc. However, you may have a look at the A330 Maintenance Manual in the appropriate section. Not sure you get the info though.

From my experience (A330 Trent engines), in many occasions, I found the autothrust very sluggish to react to speed variations, in different phases of flight (CRZ, APP).
This is very true on APP with AP off in turbulent conditions (referring to Cayenne incident). Having said that, this is a known fact and it's been documented by Airbus.

If not done already, you may also be interested in Airbus FCOM Bulletin 810/1A (JUN 04) dealing with "Thrust acceleration in A/THR modes". This applies to A320 family but it contains very good info on your query.

[Hyperveloce]

Here is a first clue about the A/THR reaction time (~4 sec.) in managed speed mode, for a given stimulis (airspeed profile) and for an A340 in approach though. In the case of Air Caraïbe the CAS dropped from 273 kts to 85 kts very rapidly, in other cases, fluctuations are superposed to the decreasing trend.
http://img182.imageshack.us/img182/938/autothrusta340cayenne.jpg

[Spooky 2]

The opening title of your thread uses the word "corrupted airspeeds". If one were to assume corrupted and unreliable were the same in this case the Boeing QRH calls for the following:

Autopilot disconnect swith Push
Autothrottle disconnect Push
Flight Director switches (both) Off
Adjust the pitch, attitude and thrustThese are memory items.

[Hyperveloce]

The following Air Safety Report (sorry, in French) shows that
1) the airspeed was underestimated to 100 kts instead of 280 kts during numerous seconds before the A/THR and A/P were automatically put off like in the AF 447 case: these "numerous seconds" would largely be enought in managed speed for the A/THR to augment the thrust level to max thrust ? (and to initiate an overspeed, to reduce the upper aerodynamic margin)
2) the pilot himself reacts to the stall alarms by pulling down the nose (and initiating a slight turn to get out from these bad conditions).
The corrupted airspeed event duration is, like in other cases, a few minutes (typically 3 or 4 mn). This enabled the crews the regain rapidly airspeeds and a normal law, but what if the AF 447 corrupted airspeed event has lasted 5, 6 or 7 minutes ?

Eurocockpit - Archives (http://www.eurocockpit.com/archives/indiv/E009479.php)