12
$\begingroup$

I saw a TV show once about Japan Airlines flight 123, in which the aft pressure bulkhead ruptured, disabling hydraulics and ripping off the vertical stabiliser. So the flight crashed after some time of flying with up-and-down oscillation.

Since then, I've also heard of two flights in which the aircraft lost control of flight control surfaces, but pilots guided their crippled plane into a landing using engine throttles only:

In principle, could such a method have been used in Japan Airlines flight 123?

In principle, would it be possible to control such a plane by use of cooperating passengers moving around in the cabin, e.g. by instruction of a pilot, to shift the weight balance of the aircraft?

$\endgroup$
  • 3
    $\begingroup$ Simon says, "Everybody run to the front of the plane!" Okay, now, "Everybody lean to the RIGHT !!!" $\endgroup$ – Tyler Durden Apr 1 '15 at 17:19
  • 3
    $\begingroup$ I'm sure one of the engine throttle only landings was United 232. What was the other one? I can't recall it. $\endgroup$ – Steve V. Apr 2 '15 at 2:48
  • 1
    $\begingroup$ @SteveV.: 2003 Baghdad DHL attempted shootdown incident, in which a cargo plane in Iraq was hit on a wing by a missile. I've edited the question $\endgroup$ – Craig McQueen Apr 3 '15 at 2:49
  • 2
    $\begingroup$ I would say that without a vertical stabilizer, their chances are close to zero due to to loss of lateral stability. I strongly suspect that the loss of the tail makes the plane's adverse yaw much worse thus making control extremely unintuitive (not sure if it's even possible to turn since any attempt to roll into a turn would make the nose point in the opposite direction) $\endgroup$ – slebetman Apr 7 '15 at 1:11
  • 1
    $\begingroup$ Magic. Most definitely magic. $\endgroup$ – Keegan Apr 22 '15 at 3:41
9
$\begingroup$

Control means to be able to vary the forces which act on an airplane, both longitudinally and vertically. Ideally, any variation will create an immediate feedback, so the pilot can "feel" how much more action is needed for the desired change. If, however, the variation of forces builds up slowly, the dynamic answer of the airplane will make the closing of the feedback loop harder to achieve. The pilot might either overcontrol and get into an oscillation, or he/she will be too careful, and the desired change never happens or happens too slowly.

Regular control surfaces are small, so they can move quickly, and their forces act on a long lever arm, so the control effect is large. Also, control surfaces change lift proportional to their deflection, so predicting the consequence of an input is easy and the feedback can be felt immediately.

Even if the primary control surfaces are ineffective, several ways of control still exist. I will list them in order of reaction speed:

The elevator can be replaced by:

  • Pitch trim,
  • Wing flaps,
  • Thrust variation, if the engines sit above or below the longitudinal axis of inertia,
  • Weight shifting, if fuel can be moved between tanks or payload can be shifted longitudinally.

The last two options work very slowly, and while they can be used to limit sink speed, it will be extremely hard to avoid overcontrol and oscillations. Situations which need precise pitch control like landings are impossible to master for an unprepared human pilot when only such crude means of control remain.

The rudder can be replaced by:

  • Rudder trim,
  • Ailerons plus pitch control in case of agile aircraft with low aspect ratio,
  • Asymmetric spoiler (if available),
  • Asymmentric thrust.

Again, the last option is rather slow and cannot be used for precise directional control.

The ailerons can be replaced by:

  • Rudder plus dihedral effect,
  • Asymmetric spoiler (if available),
  • Sideways fuel pumping.

Here again the last option is too crude and too slow for maneuvering.

If the aircraft is otherwise undamaged, and the flight dynamics are well known, a specially programmed autopilot could fly an aircraft even when using the last listed options, because it can much better predict than a human pilot how much effect its actions will have. Humans rely on a feedback loop for control, and this is not working anymore if the reaction time of the system is far above its eigenfrequencies of motion.

Note the "if": How likely will it be that only all control surface actuators have failed, but the aircraft is intact otherwise? In most cases, the control failure is a consequence of another failure which will alter the dynamic response such that an autopilot would be unprepared. There are experimental autopilots which can adapt, but they are only used by the military so far.

With the loss of most of its tail section, JAL123 would have been hard to control even if the equivalent of full rudder authority would have been given to the pilots. With thrust only, fighting the inevitable dutch roll motion by hand is impossible.

$\endgroup$
  • $\begingroup$ Note that if you have a total failure of the primary flight controls, this likely indicates a total failure of the plane's hydraulics, in which case you won't have any trim, flaps, or spoilers to work with, either - you'd be limited to fuel transfer and throttle steering. $\endgroup$ – Sean Apr 29 '18 at 18:42
  • $\begingroup$ @Sean Except for cases where trim is electric. That is not so uncommon. $\endgroup$ – Peter Kämpf Apr 29 '18 at 20:35
12
$\begingroup$

As far as I understand, the JAL123 pilots did use thrust manipulation in an attempt to control their crippled aircraft -- however, Japan's largely mountainous terrain conspired against them, as trying to figure out fly-by-thrust-lever requires a massive quantity of airspace, and unfortunately for JAL123, there was a mountain in the way of that effort. UAL232 and OO-DLL, on the other hand, both happened over relatively flat terrain, which meant that the aircrews could concentrate on working with their crippled airplanes to get them back on the ground as safely as possible, without having to worry about terrain clearance.

Of course, the missing vertical stab didn't help them either, as it meant they had to do much more work to control their airplane.

$\endgroup$
  • 7
    $\begingroup$ The missing tail on JAL123 would have severely affected the aircraft's stability. I doubt if anyone can land (or crash land) on a runway using throttles only in that airplane, even if there were no mountains in the vicinity. $\endgroup$ – kevin Apr 1 '15 at 5:07
  • 4
    $\begingroup$ It should be mentioned that as a part of the investigation, pilots tried simulating JAL123's last 32 minutes: No one could keep the plane in the air for the full 32 minutes. $\endgroup$ – Jonas G. Drange Apr 1 '15 at 12:31
  • 1
    $\begingroup$ Actually, the timing is also really important: Pilots became aware that total loss of controls could happen, so some started training for it after the incident, thus giving better chance of success after JAL123 $\endgroup$ – Antzi Apr 1 '15 at 15:53
  • $\begingroup$ I don't know the precise directional-stability characteristics of the 747, but keep in mind that, for some aircraft at least, losing the vertical stabiliser does not always necessarily preclude a safe landing (or "landing"). $\endgroup$ – Sean Apr 29 '18 at 18:51
7
$\begingroup$

As others have stated, it is possible to control a plane using throttles alone, assuming the control surfaces are largely neutral and the plane is at an stable attitude (e.g. not spinning or stalling):

  • To climb, increase throttle of all engines
  • To descent, decrease throttle of all engines
  • To turn, decrease throttle of engines on the inner side of the turn

Such feat would be very difficult as flight crews receive no training of such control methods.


On a large airliner, shifting weights would produce little effect, since the weight of passengers contribute only a very small portion of the entire weight.


NASA has already shown that it is technically possible to develop a system to automatically control a plane using throttles. Not only does the system provides stable and acceptable approaches and landings, it was also tested in unusual attitudes and at speeds 100 knots above approach speeds.

As to why this kind of system is not implemented across commercial airliners, the usual answers apply...

$\endgroup$
  • 1
    $\begingroup$ It's also debatable whether such a system could be implemented a typical airline transport jet. Military jets practically fly on thrust alone (theaviationist.com/2014/09/15/f-15-lands-with-one-wing). Airliners still depend very much on their wings, ailerons, elevators, and rudder for stability. Although UAL232 provides some anecdotal evidence that it might be possible, a lot would depend on the nature of the situation causing the emergency. $\endgroup$ – Calphool Apr 1 '15 at 15:21
  • 1
    $\begingroup$ Don't put down the NASA report that easily. The F15 considered has its engines close together. That means it's harder to turn as the torque arm is a lot shorter. The F15 also predates thrust vectoring, so it's similar to commercial planes in that respect as well. But yes, you do need two wings - that's not waht this question was about. A wing usually isn't considered as a flight control surface. $\endgroup$ – MSalters Apr 2 '15 at 13:29
  • $\begingroup$ Airliners have got into trouble from too many passengers rushing to one side or the other to watch something happening there. Things like a volcano erupting, a space shuttle launch. Rarely something that can't be corrected for, but if the normal controls are dead... I'd guess if done deliberately it might be used as a control system, but it'd be hard to control. $\endgroup$ – jwenting Apr 3 '15 at 4:37
  • $\begingroup$ @MSalters I wasn't really "putting down the NASA report", just saying that the nature of the emergency would govern the effectiveness of trying to control an airliner with thrust alone. It might be one of those situations where while technically possible, the benefit may not be there given the fact that the thrust ratios are way different between the two types of plane, and the importance of the control surfaces differs significantly as a result. $\endgroup$ – Calphool Apr 6 '15 at 20:48
4
$\begingroup$

The question here is very much related to the way in which the controls were to stop working. If the surfaces are in a more or less neutral position it would be possible to control the airplane (to an extent) using the engines based on the principal that more thrust would increase speed and create more lift as well as less thrust creating a decent. You could to some extent use differential engine power to turn as well.

In principle, would it be possible to control such a plane by use of cooperating passengers moving around in the cabin, e.g. by instruction of a pilot, to shift the weight balance of the aircraft?

Maybe, this depends largely on the plane but basic aircraft training would say yes. This is largely why you have to do weight and balance when you learn to fly. The issue here would gingerly moving everyone around so as not to put the plane in a stall.

Keep in mind that with Japan Air 123 the vertical stabilizer being blown off was a big issue...

$\endgroup$
  • 1
    $\begingroup$ On a passenger airliner, though, passengers are a rather small percentage of the overall weight. I would think they could probably make some difference in pitch, but I would think any difference in roll would be negligible, since the arm length would be rather short, especially when compared to the arm length associated with the weight of the fuel in the wings (and the engines, if wing-mounted.) $\endgroup$ – reirab Apr 2 '15 at 16:53
  • $\begingroup$ yes this is true and I would agree. But it has been noted here that passengers are negligible when it comes to weight. So it might not have an effect anywhere. $\endgroup$ – Dave Apr 2 '15 at 17:51
  • $\begingroup$ If the control surfaces fail due to the hydraulic lines/control cables being severed (rather than, say, jammed actuators), as with JAL123 (well, most of the control surfaces on JAL123; the rudder stopped working by dint of no longer being attached to the aircraft), the aerodynamic loads on the surfaces should force them back to "a more or less neutral position". (1\2) $\endgroup$ – Sean Apr 29 '18 at 18:57
  • $\begingroup$ With one exception: on some aircraft, under at least some flight conditions, the aerodynamic loads on the elevator might not be enough to keep it from dropping down under its own weight, in which case you're basically screwed. $\endgroup$ – Sean Apr 29 '18 at 18:59
0
$\begingroup$

Fly-by-wire control systems may be quadruplexed such that there are 4 channels. This is all a part of the redundancy that is built into aircraft.

Also, take a look at this article about the A380 and redundancy. http://www.roger-wilco.net/would-i-fly-on-an-airbus-a380/

A380 redundancy

A380 redundancy

$\endgroup$
  • $\begingroup$ What are those boxes on the right side of the fuselage powered by the green system? $\endgroup$ – Sean Mar 13 at 22:57

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.