Originally Posted by
CVividasku
No indeed. There is no PIO before the tail strike. But the triangle shape in the elevator deflections is a very clear indication of PIO.
The PIO appeared because the aircraft was slow to react, which prompted the pilot flying to maintain too big of an order, leading the aircraft to overreact, and himself to overcorrect. This puts too much demand in elevator deflection
speed, which is the slope of the curve. When the curve is a steadily descending and ascending line, in a triangle shape, it means you reached the elevator maximum deflection speed, which is a very clear indication of PIO. It didn't last for long as only a few triangles are visible. But it's a very clear sign. You won't find anything similar on the curves for a normal go around.
Reaching maximum elevator speed means the elevators are chasing a target that they can't reach. It means that the aircraft cannot respond timely to the pilot's orders.
As for the long landing, you can find the recordings on the youtube video just below your post... I'm not going to write the exact same thing that I already wrote. Just wait for the final report if you don't believe me. Or look for another similar incident report if you're impatient.
If you look at the curves, it's not what happens. The engines are still almost at idle when the tail strikes the runway.
So the elevators are responsible for the pitch up. If you use an image processing software to put the two curves, pilot input and elevator response, in relation to each other, you will see that there is a delayed response from the elevators, of almost one second. Then, while the elevators are not following the order they're given, the pilot waits for them while maintaining his input... If the elevators had responded, surely the pilot, who is not incompetent, would have given a smaller overall input. and he might not have stricken the tail.
Yes. It should be. But if you look at the curves, again by superposing both, you will see that if most of the time the elevators follow closely the input, in what is believably a direct law, they do not follow properly between 2 and 4 seconds before the tailstrike.
LHR is not caused by PIO. (in all I'm writing I'm not talking about YYZ because I don't have any curves obviously)
However, the fact that PIO occurred shows that the aircraft wasn't responding as expected by the pilot. And if you look closely in this 2-4 seconds before event timeframe, you will see that the elevators indeed responded with a delay.
Before saying things like this, you should at least go in a simulator, perform a go around. Then, when the curve analysis will have shown that you successfully performed a go around in spite of a one second delay in elevator response, you will be in a better position to criticise... If the elevator delay does not occur, you've proven nothing, and you're left with looking for why the elevators responded nicely in one case and late in another case.
PIO occurs easily when the aircraft response is approximately one second. Because it is also the human response time.
Have you ever flown a delta wing microlight ? Like this one :
https://images.virginexperiencedays....ompress,format
Especially in roll, these things are a bit difficult to control, at least at first. Inducing a roll requires a large force in your arms, so it takes time to move it. Then it's not ailerons that you're moving but the entire fuselage under the wing. So it's even longer to get a roll angle and a turn. Since there is a delay, it's easy especially if you over correct to get into PIO.
Indeed, the aircraft will not enter oscillations by itself, because it's dynamically stable. Dynamically unstable aircraft would be very difficult to fly and not certified.
What I'm saying since the beginning is contained here :
It follows that control system dynamics as well as airframe and pilot dynamics enter into this phenomenon
Elevator maximum speed is part of control system dynamics 
If you'd like to go into more details you can send me a PM
You called this a PIO and gave the G- rego tailscrape at LHR as an exemplar of the same. Neither of them are caused by a PIO/APIO. Now, if you provide evidence that Gz and pitch rate gets out of sync with the SSC and elevator deflection then it might be of interest. As some background, I provided expert witness testimony on PIO on an A320, so it is a subject that is of some interest. In that case there was PIO, however there was also limit cycle divergence in an attitude axis with the autopilot controlling the aircraft. In that case, the regulator, manufacturer and the accident investigation board did not look great in the courtroom. If this involved a PIO, I would be particularly interested in that. Within my investigations of the usual Monday morning funnies, PIOs were indeed noted in some events, for the electrons they invariably involved the PF not using the arm rest properly, and encountering a sharp Gz change at the pilot station, which can be from a vertical acceleration, or by static stability doing its thang, and rotating the attitude. The bus is pretty darn good however, when the schedule control deflects are correct for the aerodynamic state. In the flare, and on the ground, as well as after takeoff, the SSC gives a simple proportional deflection for the elevator deflection, (Boring added a pitch protection that reduced the elevator deflection where high pitch rates and high pitch attitudes happen on the ground (? + t buffer?), for the B773ER). As much fun as it is to evaluate control derivatives and the stability of the aircraft, looking at the longitudinal differential equations, transfer functions and Laplace transformations, they are redundant when the prong is pulled back to the stops, the plane is on the ground and the aircraft gets a cut-n-polish of its nether regions.
The following though is kind of nice as it avoids the pesky thrust couple which gets to be added the mix.
Don't get me wrong, having full backstick and then having a rapid longitudinal acceleration as the blenders come on song, increasing elevator effectiveness as a function of
V0^2, and getting the additional whammy of the pitch couple from said thrust, even as a Net outcome,
FN=((P19.V19)+(P8.V8))-(P2.V0) and that thus having a vertical offset from the CG of the bus, it adds a complexity to the control needed by the person holding the prong. As complex as that becomes if the seat occupier has gone to a full backtick, it does not constitute a PIO, it is just a really poor piloting practice. The guys that do well in such a case are often helicopter pilots, as they are trained in dealing with weird changes in control authority, stability and the whole inertia-aerodynamics deal that Locke number speaks to. Even then, helicopters are easier to manage when the variables are reduced to human manageable levels.
Wild ride PIOs are impressive, google the first inadvertent flight of the YF-16, and the Saab JAS-39 Gripen. Minor PIO is often observed in first sessions of sim on the bus,
and were a feature of the F-16 first up, which softened with the introduction of some flex to the side stick. We see a form of PIO with the ATR-72 landings sometimes, that introduces additional forces into the body response, enough to be interesting. The solution in most go these cases is to remove the P from the PIO, just momentarily is enough normally. The gain and lag in the control loop are the other variables to be played with to stop getting ugiies.
Tin Hat time: One event that I investigated which was pretty funny was attempted to be blamed as a PIO due to the OEMs control system. The driver hit the ground hard in his trusty 2 holer, second hard landing in 2 sectors. He/she remembered the no fault position of the aircraft operator where a GA was conducted, and so decided to GA, yelling out, "GoAround"... and pulled back on the prong. Unfortunately, he had already selected reverse, and tapping (pulling in this case) the TOGA paddles didn't increase the noise, both engines were obediently in reverse. The speed was still high enough for the boys room to be elevated, and the tail fortunately stopped the plane from tipping backward too far. After some time, a long time, with the plain still refusing to get airborne, the occupants of the flight deck realised the plane was decelerating, and realised they were still in reverse. Luckily, they got both engines to cancel reverse, and then bent the thrust levers forward mandraulicly. The attitude remained nice and stable, more or les, only increasing slightly as more and more bits o' plain got wored offn'' it. Eventually, the plain rejected planet earth and got airborne, however, it did perform a little better as the OEW was reduced by the tail cone ,APU, paint and sundry items that traveled the length of the runway behind our intepid aviators. The bolder plane got cleared for a circuit, much to the surprise of all concerned in the investigation, and as a sister ship had observed the debris being dropped on the runway and had advised tower, the lads landed on the parallel without further incident. On approaching the gate, the FO found he couldn't get the APU to start, so on chocks external power was applied. Plane signed off as APU INOP. By the time the crew got ready to depart the pointy end, there was quite a congregation of engineers who had seen a lot of stuff still falling off the back of the plane at its leisure. There have been many impressive tailstrikes, but this one was in a league of its own, bad enough that the aircraft never returned to service. In the investigation, there was a political line towards blaming the OEM for the "PIO" that had occurred, and not only did I object to that, so did the test pilots from the OEM. Remedial training was given and our recommendations on GA were taken up.
LONGITUDINAL CONTROL stuff
While the above equations give a nice way of describing and modelling the control laws, and are the right ones to be considering in this non PIO case, the problem arises when the pilots inputs get out of sync with the control system or aircraft response. That is the differential of the inputs and the responses. Humans approximate a PID controller if they are well trained, otherwise they are barely proportional, they don't integrate the response without training, and they don't do the differential bit without experience in the response of that system. Airbus's own control system (C*, which is C*= n + A*q) was in effect a PI controller, and there was some consideration on their enhancement to PID which may well have occurred by now, (I haven't messed with an A320/330 or 340 for a few years).
PID stuff
[Engine acceleration is non linear, it is.... a third order function with two asymptotes, one at each end. If the controller is hydromechanical. then it tends to look like a 5th order function with little wiggly bits at the max thrust limit. Thrust increase is itself a 2nd order function, it follows a rather common curve].