11
$\begingroup$

I've always been fascinated with the impact of ailerons on an aircraft's trajectory; from what I have seen sitting by a wing is that the slightest movement can bank the aircraft quite significantly.

On my last two flights we had mild to moderate turbulence and I noticed that nearly every movement appeared to be counteracted. That is, if the plane moved to the left, it was (in second to sub-second time) accompanied by a sensation of moving to the right. It's the same for up and down.

What I couldn't decide and what I can't seem to research properly is between these options:

  1. Given the small adjustments needed on flaps, and the speed of software and hydraulics these days, the aircraft is actively correcting for every small disturbance it can, similar to traction control in cars
  2. The aircraft design is just self-righting, similar to a gyroscope. The fact that a left movement seems to be accompanied by a right movement is just the fact that the force pushing the aircraft left was fleeting and the aircraft just abruptly stops being forced left, giving me the sensation that I then pulled to the right
  3. I'm a human being trying to find patterns in something and, actually, the plane is just being buffeted randomly

Is the aircraft actively responding in real time to these minor disturbances?

$\endgroup$
1
  • $\begingroup$ My main curiosity is just whether a commercial airliner is actually making adjustments in fractions of a second in mild turbulence or whether it was just the property of the plane design itself (point 2). I perhaps presented it incorrectly in the question. The answer seems to suggest that indeed the aircraft really is as responsive as what I thought I felt $\endgroup$
    – roganjosh
    Sep 28 at 17:56

5 Answers 5

16
$\begingroup$

Stability by design

Transport aircraft are inherently stable. They are designed with a positive static stability, meaning they tend to automatically correct for variations in their attitude and return to neutral. E.g., the wings are not built horizontal; they make an angle (dihedral or anhedral) to return to zero bank after a wind gust. Without this roll stability, the aircraft could continue to roll after this motion has started.

However, this stability mustn't make piloting difficult, e.g., too much roll stability would require excessively large efforts on control surfaces. The compromise is a design choice, e.g., fighter aircraft have less stability in order to facilitate fast maneuvers, while transport aircraft are less agile, but more stable.

With less stability, safe piloting may require computers (e.g. to move canards). The same principle of aided flight is also used on more stable aircraft, to relieve the pilot from constantly adjusting attitude, or to react faster than the pilot, e.g. for gust compensation.

When you feel instantaneous reaction, it's likely an action of these computers.

Flight computers for additional stability and protection

Additional stability isn't natural, but managed by computers distinct from the autopilot maintaining programmed route and altitude, dedicated to the actuation of control surfaces. This system is what is generally called fly by wire.

Computers receive pilot orders, calculate a combination of surface movements, within predefined safe limits, send electrical signals to actuators, ensure the calculated changes are carried out, then maintain the commanded attitude.

For attitude keeping, the principle is to detect accelerations and orientation changes on each of the three axes. This is usually done by the inertial reference system which also computes the aircraft position. These inertial data are used by the flight computers.

One example of such computer is the yaw damper, which counteracts both a yaw and roll oscillation in order to prevent a flight dynamic mode known as Dutch roll which can be catastrophic if left uncontrolled, and has been the cause of accidents.

Airbus example

On Airbus aircraft, attitude correction is controlled by seven computers (ELAC, SEC, and FAC, see details in Airbus control architecture - where does the actual autopilot live?) which can reconfigure themselves if one or more fail, to continue providing protection and stability as much as possible. They work regardless of who pilots, human or autopilot.

enter image description here

These computers set the attitude commanded by the pilot, and maintain it when input has ceased. They also ensure the commanded attitude and the commanded rate of change are within the possibilities of the aircraft (flight envelope protection). On the schematics above there are two mechanical links which are last resort manual controls to land the aircraft in case the fly by wire system has failed to the point orders cannot be transmitted.

For bank stability, this is the way Airbus A320 family works. From manufacturer FCOM:

Enter image description here

and:

Enter image description here

$\endgroup$
13
$\begingroup$

When the autopilot is engaged, which is typical during cruise flight, it's working to hold the wings level and maintain the desired pitch most of the time (a programmed turn or climb/descent to follow the navigation being the exception). So it detects any deviation from wings-level or desired pitch and makes a control input to promptly return to the desired attitude.

You see that control input with the aileron and spoiler deflection; you don't see the elevator inputs for pitch control since that's behind the cabin.

A human pilot flying the airplane does much the same thing, perhaps not quite as quickly as the autopilot.

If neither the autopilot nor the human pilot were providing the corrective inputs, then you wouldn't see the ailerons moving, and the aircraft would spend far longer in an attitude other than wings-level, and probably wouldn't hold altitude very well.

$\endgroup$
7
  • $\begingroup$ Can the autopilot really counteract the motion of turbulence within a second? That's what blows my mind if theory (1) is true. $\endgroup$
    – roganjosh
    Sep 28 at 14:14
  • 13
    $\begingroup$ Within one second? Sure. Consider driving on the highway in gusty winds... you react way faster than 1 second when a gust starts to push the car out of the lane, and digital autopilots can recognize an acceleration in the roll axis even faster than we can perceive it. $\endgroup$
    – Ralph J
    Sep 28 at 14:18
  • $\begingroup$ I love the way you say that so casually when we're talking about several hundred tonnes being moved around without the traction that a road would give you in the car analogy! :) $\endgroup$
    – roganjosh
    Sep 28 at 14:21
  • 5
    $\begingroup$ Fair point... the car analogy is more about how fast the roll can be perceived; the sensitivity of the controls is a different topic. Airliner flight controls are not as sensitive as those of acrobatic aircraft, where a little push can send you rolling over & over! But, they do have ample authority to do all that's needed to keep the airplane pointing where it needs to go, and the autopilot is tuned to command the right amount of deflection - powered by 3,000 psi hydraulic systems - to arrest the undesired roll & return to level flight promptly & smoothly. $\endgroup$
    – Ralph J
    Sep 28 at 14:29
  • 1
    $\begingroup$ @AnonymousPhysicist Added comments for pitch, though the same basic principles apply. Light to moderate turbulence doesn't tend to induce very much yaw, at least with a yaw damper engaged (which it is full-time, if everything is working as normal). $\endgroup$
    – Ralph J
    Sep 28 at 22:31
11
$\begingroup$

I suspect you were flying on an aircraft with a gust/load alleviation system such as the Boeing 787, Airbus A350 or A380.

In these aircraft, accelerations due to turbulence are measured in the wing and immediately counteracted by deflecting the control surfaces.

This works independently from the autopilot.

It sometimes gives the impression that the controls are somewhat randomly fluttering in the wind.

From this answer:

A gust alleviation system is a control system fitted to some Fly-by-wire (FBW) aircraft that reduces the effect of gust loads on the aircraft by deflecting control surfaces such as ailerons, rudder and elevators.

The system works by measuring the upward acceleration of the aircraft and comparing that with the acceleration commanded by the (auto)pilot. A feedback loop adds a correction signal to the signals controlling the deflection of the control surfaces in order to counteract the accelerations cause by wind gusts.

In its simplest form the accelerations are sensed near the center of gravity of the aircraft. More advanced implementations work with multiple sensors in the body and in the wings. These gust alleviation systems do not only attempt to annul the effect of the gusts on the aircraft's normal (upward) acceleration, they also reduce the wing bending moments. This in turn reduces metal fatigue.

In the A380 the load alleviation system is nicknamed "valse des aileron" (Waltz of the Ailerons) because of the dancing movements of the aileron:

See also here

$\endgroup$
4
  • 2
    $\begingroup$ Since the wing is elastic "measured in the wing" probably means the response occurs well before the passenger gets shaken. I suspect these gust alleviation systems work so quickly that they do not contribute to the passenger's feeling of a "response." I think they just reduce the amount of movement the passenger feels. Do you have any more information? $\endgroup$ Sep 28 at 21:04
  • $\begingroup$ @AnonymousPhysicist they certainly reduce the amount of movement the passengers feel, but there are still residual forces transferred from the wing to the fuselage. So while the start of the response may take place before the elastic wave through the wing has reached the fuselage (but this is not necessarily the case), the wave head will reach the fuselage. The GAS reduces the magnitude of the wave. $\endgroup$
    – DeltaLima
    Sep 29 at 7:01
  • $\begingroup$ Gusts can also be detected at the nose, before they reach the wing. At 500 kt this would leave about 0.1 s for the actuators to move ailerons and spoilers, when gusts are ahead. $\endgroup$
    – mins
    Sep 29 at 16:54
  • $\begingroup$ Most Airbus models have a turbulence damping function. The early A320 had it but it had technical issues and was dropped. The sharklet models got it back due to wing loading. A330/A340 had it due to fuselage length, it dampened flex modes for fatigue life and comfort, it was a fundamental benefit of FBW. $\endgroup$
    – user71659
    Sep 29 at 17:42
4
$\begingroup$

That is, if the plane moved to the left, it was (in second to sub-second time) accompanied by a sensation of moving to the right. It's the same for up and down.

Human perception is a bit strange. When you feel the plane move "up and down," you are not feeling its position. You are feeling acceleration. If:

  1. The plane is flying at constant altitude.
  2. You hit some turbulence and feel and "upward movement", then the plane is accelerating up.
  3. The plane's altitude increases.
  4. Very soon, you feel an equal "downward movement", then the plane is decelerating up.
  5. If the two accelerations are equal (but opposite), then the plane is flying at a constant altitude, but now it is higher up.

Yes, the aircraft autopilot will return the aircraft to the selected altitude. However, that's probably not what you are feeling.

I'm a human being trying to find patterns in something and, actually, the plane is just being buffeted randomly

That is how humans normally work. If you take a bunch of random buffets, most of them are going to cancel out quickly.

Airplanes (and particularly large modern airliners) are not rigid. You need to consider the bending of the aircraft as well as the buffeting to understand the passenger experience.

$\endgroup$
3
  • $\begingroup$ efcms.engr.utk.edu/ef151-2019-01/m1/pilot/class15/… $\endgroup$ Sep 28 at 21:07
  • $\begingroup$ Given the other answers, I'm not convinced this was an issue of perception. The fact that I was aware of my own fallibility at the time when I was trying to tell whether there really was an effect, and a background in engineering, I tried really hard to remove that bias. I don't recall feeling "corrective force" in flights I've had in the past so this felt quite distinct as direct opposition to buffeting. Of course, it's hard to tell when turbulence is... turbulent and you don't get clean 1:1 signals since it's happening almost continuously $\endgroup$
    – roganjosh
    Sep 29 at 12:10
  • $\begingroup$ @roganjosh If you are saying you observed different behavior on certain flights, then I don't think my answer will help you. Maybe it was a different type of aircraft. Maybe it was different turbulence. $\endgroup$ Sep 29 at 13:02
1
$\begingroup$

Do airliners respond to mild turbulence in real time?

More or less.

Gust alleviation is a very important aspect in aircraft design. It is so important that big companies have a whole department only devoted to gust studies. For a jetliner entering a gust, the relevant loads are indeed the most critical loads normally encountered in its life and therefore its structure is designed around those loads. A reduction of those loads leads to savings in weight and fuel as high as 5 to 10%.

The following picture (source) shows a typical distribution of gust loads on a jetliner wing:

 gust loads

Especially important is the increase of the aerodynamic loads towards the tips of the wing since it generates a high bending moment at the wing roots.


There are many methods to counteract those gust loads but they can be classified either as passive or active according to their working principle.

A passive method can be for example to design the structure of the wing in such a way that when it bends upwards, the local AoA gets reduced reducing the local aerodynamic loads and therefore the bending moment.

A standard active method implies instead the use of gust sensors feed in an automatic system which uses some control surfaces to counteract the load increase. In this case, ailerons are typically used since they are already located at the wing tips i.e. where the highest gust loads happen. Using ailerons, spoilers or flaps is convenient also because they are already there i.e. there's no need to add ad-hoc surfaces with their relevant weights. Anyway, as you suspected, the whole loop (from when the sensors pick up a gust till the ailerons get moved) is not as fast as the whole gust spectrum, which ranges from few to hundreds of oscillations per second. Being ailerons, spoilers and flaps mainly sized for manoeuvring and high-lift devices, they are large and heavy and hydraulically actuated. They are therefore quite slow and virtually ineffective against the high-frequency part of the gust.

To gain some precious milliseconds, Airbus tested some time ago a "high-resolution direct-measuring short-pulse ultraviolet (UV) Doppler lidar system". This system detected gust movements till some 50m in front of the airplane giving 300ms lead time to the gust control system. I don't know if this particular system made it out of the experimental phase.

$\endgroup$
2
  • $\begingroup$ That thesis is bang on the nose as I'm reading through the introduction. Thank you for adding that $\endgroup$
    – roganjosh
    Oct 3 at 18:44
  • $\begingroup$ @roganjosh: you're welcome 🤗 $\endgroup$
    – sophit
    Oct 3 at 19:39

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .