Jet Upset Training - For or Against.

[Centaurus]

There are differing opinions on the perceived value of unusual attitude recovery training in simulators. Personally I am all for it, particularly having seen experienced airline pilots who should know better, not having a clue how to recover to right way up after being placed inverted during simulator training.

The argument against unusual attitude recovery training is that the simulator does not replicate the real thing. Control forces may be different and aircraft behaviour unknown, because no test pilot is going to risk an aircraft in such extreme manoeuvres.

Some advise that the next best thing is to learn aerobatics in a flying school aircraft and use those techniques where necessary to recover in a Boeing type for example. But again we are faced with the same argument as before and that is the Boeing type may not react on recovery in the same way as a light aircraft would.

Accepting the conservative premise that extreme unusual attitude recoveries should not be practiced in the simulator for fear of getting the wrong gen, then do we close our eyes and say it will never happen to me and thus no need to be prepared?

Or do we accept our gut feeling that yes - you can barrel roll a 707 successfully (ever see the photo of the Boeing test pilot way back in the Fifties doing a demo barrel roll on the prototype 707?) and you could do the same thing in a 737 too (maybe?) and get away with it. So very probably you could barrel roll the 737 in a simulator and recover very nicely, thank you very much. And its a fair bet that the real thing would be almost the same as the simulator.

That said, surely the value of the simulator in training for unusual attitude recoveries is the knowledge of the flight instrument indications when in an extreme attitude such as full inverted (behind the wake turbulence of an A380 maybe).

A barrel roll in a simulator is not only a good lark - but there can be serious lessons learned here in sky-pointer positioning, rate of roll recovery, basic flight instrument indications, rate of speed decay, increase and so on. There must be more value in this than learning jet upset recovery from a picture book or theory manual.

If one accepts that in real life the control forces on recovery may be different, and that what matters is the quick identication of the position you have found yourself in, and which way is the quickest to get wings level, then we should not ignore the importance of unusual attitude recovery training in the simulator - despite it's limitations.

If aerobatic practice in a light aircraft is considered good training for jet upset recovery notwithstanding control force differences, why not the same principle for the flight simulator using the actual flight instruments as in the Boeing etc?

[GlueBall]

not having a clue how to recover to right way up after being placed inverted during simulator training.

Captain Centaurus: The real world and the simulator are 2 different animals.

Captain Centaurus: Why don't you spend some time in a 74 sim and see if you can actually achieve the physical sensation of getting yourself inverted...then come back and tell us.

[Chimbu chuckles]

I'll have to try the cuban 8 in the 767 next sim recurrent.

The 767 rolls beautifully as long as you're over 300 IAS...my party trick in the 767 sim is a aileron roll started at 50 RA and then a climbing steep turn, derry roll, flight idle then back around for a glide approach and landing.:E

Whether the real aeroplane could do as well is irrelevant...it's fun.

As far as upset recover is concerned the point is not necesarily about whether you can or can't simulate the G it's about recognising what is happening and using the controls as required to achieve brown side down without ripping the wings off or doing a supersonic lawn dart.

I personally think the 767, 737 are perfectly capable of being rolled...and I think the control responses when rolling the sim are reasonably realistic.

Chuck.

[411A]

AA tried the 'upset recovery' technique with rather poor results........AA587 into the NJ shore.
296 dead.

Next question?

[Kaptin M]

Recovery from unusual attitudes has been a part of sim training of all (5) airlines for which I've flown, and I personally believe there is some value in them.
The object of the exercises is to ensure that the pf correctly recognises and applies immediate recovery technique.

The accident you have cited 411A had a lot to do with (i) the aircaraft being an Airbus, (ii) having a "plastic" vertical stabiliser, and (iii) the pf being "over zealous" in his application of rudder.

[Centaurus]

Love the humour, Glueball. Especially your profile.

[john_tullamarine]

I am right with Centaurus et al in the sensible exploitation of the sim's training capability.

Sure, the sim is not an aeroplane and only goes so far in "simulating" flight. However, is it not reasonable to use it to develop generic skills, such as UA recovery, even if the real thing is going to be different to a greater or lesser extent ?

On another tack, I have seen many pilots experience a significant confidence boost after some time playing with sim UAs (normally with the motion off - the main goal is the I/F value). Barrel rolls are fun - could never get the 737 sims to do a loop - but the value is in repetitive exposure to odd UAs .. banked through the vertical nose high/low type exercises.

Just not doing it because the sim isn't the "real" thing is a poor rationale and akin to burying one's head in the sand. The situation doesn't happen in the real world all that often to big aeroplanes but, when and if it does, there's not all that much time to waste before the situation becomes irrecoverable.

I don't see as much value in using a light aerobatic trainer due to the very different control loadings and inertial response characteristics .. not to suggest that one shouldn't play aeros in trainers .. but in addition rather than instead of sim work.

[Chimbu chuckles]

411A I would suggest that the AA crew tried a upset recovery technique rather than the upset recovery technique.

Perhaps better training in the sim might have mitigated that disaster...whether or not it was a Boeing or that French piece of ****e:E:} :p

[411A]

<<Perhaps better training in the sim might have mitigated that disaster...>>>

Would agree completely.:sad:

[Charles Darwin]

being an Airbus, (ii) having a "plastic" vertical stabiliser,

I think the 787 (that´s Boeing) is going to be all plastic...except the rudder!:E

[alf5071h]

A major contribution to upset incidents is the surprising onset of the event. Simulators can provide valuable training that improves crew situation awareness that may mitigate surprise or even a loss of control. This is also the time and place to control any adverse behaviours for acting without assessment, rush, or carefree responses. However simulators have many limitations when attempting to simulate the motions and sensations of an upset. They will not for example simulate the false climb illusion or anything similar that may be encountered during an upset or the recovery.

There are few really gross upsets; one that can be simulated and opens many eyes is fuel imbalance. With the autopilot engaged in level flight, and unbeknown to the crew, set the maximum lateral fuel imbalance before any alert is given. Disengage the autopilot – the surprise; and then assess how the crew use the cues of motion, instruments, and control feel to complete their recovery – if at all. The lesson to be learned is that the aircraft may not be in trim (zero force may not equal stick centre) and that the instruments are the most important cue.
For those doubters who say that this scenario will never be encountered, ask the UK 737 operator who had just such an event – at least 120 deg roll. Also, with some speculation, did a lateral force imbalance contribute to the recent Middle East 737 accident due to a possible slat asymmetry?

Thus the most important cues for upset recovery are obtained from the instruments, always believe the instruments; this too is the defence against spatial disorientation. Aerobatic training in aircraft or the simulator may give personal confidence in the aircraft’s capabilities, but it does not provide upset recovery training. This is achieved by dedicated instruction and practice of how to recover from unusual attitudes by use of instruments, first to asses the aircraft’s position, motion, and trends; then with rules and guidelines, safely recover to controlled flight as quickly and as safely as possible.

[redsnail]

The company I did my initial endorsement for the Hawker (yes I realise that it isn't in the heavy category) used a scenario that was most likely to set up a jet upset.

They would use wake turbulence from a heavy overflying to set it up. We'd get a TCAS alert (and have to respond) and then get the jet upset. First time I had seen UA's set up like that and I thought it was very effective.

[Centaurus]

Must say I am personally delighted at all the replies so far. My posts invariably result in a storm of derision.

Next subject is the practice at GPWS recoveries in the simulator. This is another sequence rarely encountered in simulator training. I am talking about running into the glass mountain scenario rather than a sink rate sink rate, glide slope glide slope event.

In my view it is a vital part of simulator training to get the GPWS manoeuvre perfect. This includes from a high speed descent with speed brake extended where the sequence of pull up while simultaneously firewalling the thrust levers, is followed lastly by retracting the speed brakes.

You may recall that the speed brakes (in the 737 at least) do not retract automatically in this type of escape manoeuvre whereas they do retract automtically in a rejected landing. A 757 was lost when the crew forgot to retract the speed brakes and although the manoeuvre was executed in very quick time, the aircraft hit the mountain just a few feet from the summit.

I have seen surprised delays when the GPWS goes off during IMC descents and the simulator took the tops of hills because of these delays. GPWS manoeuvers, like unusual attitude recovery practice, is often seen as a fun game thing, rather than serious instruction.

[ZFT]

One of the problems associated with EGPWS is the difficulty facing the simulator operators in introducing ‘glass mountains’ as the Honeywell terrain (and obstacle) database is embedded within the EGPWS computer.

[alf5071h]

ZFT the solution to your problem of introducing ‘a glass mountain’ is within the EGPWS box. As you state the system has an internal terrain database, and assuming that your simulator uses the real EGPWS hardware (the most sensible solution), then the terrain profiles can be accessed by the simulator - a real world terrain database for free. If you have any problems, and there could be many with the poor standards of simulating aircraft geographic position, then contact Honeywell as they have helped many operators. EGPWS Simulators. (http://egpws.com/general_information/simulator_information.htm)

The use of the internal EGPWS terrain database in simulation can adds some interest to check rides; just how close do you fly to terrain during an engine out procedure. One operator had a nasty (positive) surprise during a simulated emergency turn procedure when the aircraft failed to clear the terrain due to incorrect performance planning.

Centaurus I totally agree with you on the need for practicing GPWS pull ups – how to handle the aircraft and to gain a feel for its capabilities during emergency procedures, to experience non normal attitudes, noting the speed loss and rate of change of parameters. In these respects the same applies to ACAS manoeuvres, how many pilots really know what a 0.75g push for a descent feels like?
With the advent of EGPWS practicing the pull up procedure is just as important, but this need suggests that earlier alerts or awareness information has been missed or ignored – which has been the case in some incidents. With a terrain amber alert the general procedure is to manoeuvre the aircraft to correct the situation, but the occurrence of the event probably indicates that the crew have misunderstood the situation, thus why not climb to a safe height and reassess. Those who do not climb / manoeuvre will increase the risk of having a red Terrain warning, which will require a well practiced pull up. With EGPWS it is important to make full use of all of the features, e.g. at least one pilot should have terrain displayed during take off and landing, and peaks mode can be used during cruise to identify mountainous terrain which may increase turbulence and the possibility of an upset. In aircraft that have the bank angle alert activated, this mode may provide that additional warning that prevents a full upset.

Good knowledge and correct use of safety equipment will aid all operations; ACAS, like EGPWS has shown many side benefits i.e. why your descent/climb clearance is denied due to preceding / crossing traffic – and it could give early awareness of possible wake turbulence.

[Charles Darwin]

I think simulator manufacturers lack data outside the normal operational envelope of the aircraft. There is no guarantee the jet will behave in the same manner that the sim, when ie. upside down og knife-edge. I´ve flown both 737 and 757 sims inverted without problems but I have no doubt the actual experience would differ somewhat. I however think some practice in the sim could not hurt in any way, but beware the inaccurate data.

[ZFT]

alf5071h

Yes we do use as aircraft EGPWS hardware and again, yes we are talking to Honeywell and I totally concur that they are very helpful.

With GPWS it was a lot more straightforward to introduce ‘glass mountains’ as the terrain was controlled by the simulator and/or visual system as opposed to the Honeywell database.

With regards your comment “poor standards of simulating aircraft geographic position”. The big problem facing the industry is that all the different nav data sources used within the simulator do not necessarily agree amongst themselves. There are geographical errors/differences even between all the principal nav data suppliers for an identical facility. This is compounded because Honeywell are reluctant to reveal the source of their EGPWS data. This problem is further compounded when constructing visual databases when their geographic data also differs from the nav data!!!!

Very interesting topic though.

[Iron City]

A problem with doing unusual attitude, stall, and uncontrolled flight training is simulators is how well the simulation represents the real world. In general, it is probably better to do training like this is the sim than to have no training at all unless the result is "negative training" in the form of giving the crew the idea that they can do something in the a/c that they in fact can not.

The sim itself can only do what it is programmed to do, and it can only be programmed with a good quality simulation when there is data available from the real aircraft. The products of the aero engineer's art provides indications of how the a/c might perform but remember that the equations describing the performance of the a/c become discontinuous at the spin entry. The predicted performance on the basis of the design are also okay only for relatively steady state configuration, speed, attitude etc, and don't tell you very well what happens over time when you make a particular control input or configuration change. In other words, don't trust the simulation the farther it gets from the tested flight envelope.

That said, I would bet that in a well built and well maintained (not tweaked all over the place) sim the barrel rolls and such would go pretty much as they would in the real aircraft.

[411A]

posted 28th April 2005 20:59
--------------------------------------------------------------------------------

A problem with doing unusual attitude, stall, and uncontrolled flight training is simulators is how well the simulation represents the real world. In general, it is probably better to do training like this is the sim than to have no training at all unless the result is "negative training" in the form of giving the crew the idea that they can do something in the a/c that they in fact can not.>>

Well now, imagine that.
I wonder if AMERICAN AIRLINES has awaken to this idea?
Stomp on the rudder...you die, (at higher speeds), clearly something that the 'young turks' in the training department never knew about.
I wonder just why?
Others have known, for a very long time.

AA...sky Gods at their 'best'.:{ :{ :yuk:

[AirRabbit]

From what I’ve read from posts by Alf5071h and Centaurus – I must agree with them both. Simulators are wonderful training, testing, and checking tools – but just like any good tool, you have to know how, when, and where to use it and, just as importantly, one must know when its not appropriate to use that specific tool. You can probably drive a Phillips head screw with the claw end of a roofing hammer, but I don’t know what you’d do to the screw, the hammer, or your patience!

Simulators are programmed with data that is gathered during flight testing of the airplane. Flight testing does not go much beyond the “normal” flight envelope of the airplane – something like 45 degrees left and right; 25-30 degrees nose up pitch, 12-15 degrees nose down pitch, and very, very small side-slip angles. Stay within those parameters, and the simulator is very likely to perform just like the airplane. Go outside of those parameters, and you’re in the middle of no-man’s land.

The good news is that if we specifically avoid discussing procedures and/or technique, we can use simulators to show pilots what the outside view of world looks like at all sorts of unusual attitudes, and we can compare that view with the instrument displays pilots look at all day. Using good common sense, we can talk about how to determine the shortest distance to roll back to level flight, what causes increased “g” loading (straight and rolling), and how easy those “g’s” add up when recovering from such an attitude. We can use operations within the validated flight envelope to demonstrate what the controls feel like when the airspeed is down around the stall. We don’t have to timidly approach that voodoo point and retreat from it like a scared rabbit. We can deliberately stay at that airspeed and maneuver the airplane (simulator) to see how it handles. We can set up high rates of sink and recover to see how the airplane (simulator) responds; how quickly it builds up “g” loading; how long it takes to arrest the descent; etc.

Yeah, teaching all the fine points of getting into and out of unusual attitudes in a simulator is NOT something we should be doing. But we don’t have to throw out the baby with the bath water.

Thanks for letting me put in my 2 cents.

AirRabbit