Came across this article tonight:

 

http://www.aviationweek.com/aw/generic/story_generic.jsp?channel=bca&id=news/bca/2012/04/01/BC_04_01_2012_p34-432588.xml&headline=Upset%20Recovery%20in%20Sims&next=0

 

Given that we semi-regularly receive tickets and forum posts from customers who expect us to have modeled stalls and things like that in perfect detail within FSX, this might be eye-opening as to just how inaccurate the multi-million dollar level D sims are with it...

Excelent article, thanks for sharing. I see this issue is still around (sims). As I remember the type of math required to solve the equations of motions for edge-of-the envelope situations was very complex (non-linear kind). On a related note, after reading the accident report for flight 447, I wondered what kind of stall recovery involves pitching up? The only way I could see using this technique is if those engines had an extraordinary amount of power availeable (assuming you still have pitch control authority). I mean this is basic flight school stuff, even with computer control, what difference does it make? Normally speaking, in a stall situation you can trade altitude for airspeed (point the nose down) to get yourself out of the stall and regain maneuvering capabilty (adding energy (power) too), unless you are a fighter jet where you can just punch your way out of trouble with thrust (assuming also you can still point the nose where you want it to go of course).

Excelent article, thanks for sharing. I see this issue is still around (sims). As I remember the type of math required to solve the equations of motions for edge-of-the envelope situations was very complex (non-linear kind). On a related note, after reading the accident report for flight 447, I wondered what kind of stall recovery involves pitching up? The only way I could see using this technique is if those engines had an extraordinary amount of power availeable (assuming you still have pitch control authority). I mean this is basic flight school stuff, even with computer control, what difference does it make? Normally speaking, in a stall situation you can trade altitude for airspeed (point the nose down) to get yourself out of the stall and regain maneuvering capabilty (adding energy (power) too), unless you are a fighter jet where you can just punch your way out of trouble with thrust (assuming also you can still point the nose where you want it to go of course).

 

Pre-AF447 stall recovery training in airliners focused on 'minimum loss of altitude'. If you read the training manuals they usually recommend a relaxation of back-pressure, not a full-on nose down command. Stall recovery training also usually assumed the pilots would react at the stall warning, i.e. before the plane is actually stalled. Finally since stall recovery is such 'basic flight school' stuff it was usually only covered during the initial type rating and not in recurrent simulator checks.

Pre-AF447 stall recovery training in airliners focused on 'minimum loss of altitude'. If you read the training manuals they usually recommend a relaxation of back-pressure, not a full-on nose down command. Stall recovery training also usually assumed the pilots would react at the stall warning, i.e. before the plane is actually stalled. Finally since stall recovery is such 'basic flight school' stuff it was usually only covered during the initial type rating and not in recurrent simulator checks.

Relaxing back pressure is more accurate for basic flight training, you are correct. I don't see how applying back pressure would be a stall recovery technique on any stall situation though. The pilots in that flight kept pulling back on the stick as the report says, until the captain finally made it to the cockpit and recomended otherwise (too late). One thing to note was that the A/C at that point was not operating in some "protected" computer assisted mode, but under another operating mode, so the recovery techniques may have been crossed. Anyway, it would seem like a confusing mess on an already stressfull situation...

Excellent read Ryan, thank you for sharing.

 

On page 4 towards the bottom:

"Flight simulators “are ‘virtual aircraft’ and they should not be used to develop techniques at the edges of the flight envelope,” Cautions Capt. Wainwright. Rather, he advises, “Concentrate everyone’s attention on taking action early enough to prevent the occurrence of loss of control.” That advice is echoed by other studies."

 

Makes perfect sense and seems logical to focus on preventing the LOC in the first place especially after knowing that LOC recovery in FFSes is seriously flawed.

Interesting stuff.

 

A lot of what people forget about airline training simulators, is that they are as much about training familiarity with switch locations and the roles and division of labour between the crew for good CRM, as they are about modeling the flight characteristics of an aircraft pefectly, although the full-motion ones do indeed attempt to at least have a good stab at simulating how the real thing flies, and that is neither perfect, nor cheap to do.

 

There are less fancy 'procedural trainers' which can be used too, in fact, some of Boeing's 787 training aids at their main training centre for pilots converting to the 787, use nothing more complex than a classroomn full of very basic office-type seats with a couple of touch screen computers located where the main panel and the pedestal are in relation to those seats, simply to get pilots used to moving their hand to the right spot to find a switch. Most home flight simmers have a set up that looks waaay better than that, but it is not necessary to have one in order to train pilots properly, so they don't waste effort on trying to make it look any fancier.

 

FS developers for home PC users have a bigger computing task on their hands than is often appreciated, since buyers of their add-ons have much higher expectation levels, and developers have to meet those expectation with something which works on a single PC, which could have one of several operating systems, and all kinds of different hardware configs, using a base code that was written seven years ago!

 

Al

Interesting, thanks!

Pre-AF447 stall recovery training in airliners focused on 'minimum loss of altitude'. If you read the training manuals they usually recommend a relaxation of back-pressure, not a full-on nose down command. Stall recovery training also usually assumed the pilots would react at the stall warning, i.e. before the plane is actually stalled. Finally since stall recovery is such 'basic flight school' stuff it was usually only covered during the initial type rating and not in recurrent simulator checks.

 

Not to mention that they're flying an aircraft that's supposedly "impossible" to stall - I think there's more than a bit of the Titanic mentality going on when they start saying things like that. I don't think those pilots believed they were really in a stall - there's no other explanation...

Airbus never says it is impossible to stall, on the contrary, thy emphasise it IS an aircraft bound by the laws of gravity.

They pioneered Alpha Floor protection many many years before any other aeroplane maker did, and the MD11 is the first that had something similar with their low speed protection.

The problem for AF447 is that the pitot tubes were frozen and the protection reverted to manual. This is were good old basic airmanship should have kicked in by the pilots and they actually made a hash of it.

A FBW Airbus is indeed impossible to stall, so long as the system law which overrides elevator inputs when you yank the stick back too far or kick on too much rudder is operational, and providing the aircraft does not lose a massive amount of headwind suddenly. But there is only one system mode on an Airbus where that level of protection occurs, and it is one of five modes that could be in operation:

 

Normal Law: this is the one where it is impossible to pitch it up past a safe angle of attack. It provides speed, pitch, bank, yaw and airframe stress protection, by limiting control authority if the stick or rudder is overused or if too much throttle is set. In other words, this is the 'no stalls mode'. And it is the only one that does that.

 

Alternate Law: this is the mode which kicks in if some multi-redundant systems fail. Some flight envelope protection is lost, in fact most of it is lost apart from stress protection limits, it is annunciated on the MFD with a warning that there is no protection from overcooking things on the controls, and so the aeroplane can indeed be stalled in this mode, and you could probably damage the aeroplane with overly big control inputs too.

 

Direct Law: this is the one which kicks in if even more systems drop out, the MFD warns about trim settings and the control surface deflections are, as the name suggests, directly related to the amount which the controls in the cockpit are moved. In other words, you could cheerfully rip the wings off if you where a clumsy sod when this mode was engaged.

 

Mechanical Law: this is the one you get if all electrics are out and you are on emergency battery power. A big red pitch trim warning shows up on the PFD to let you know things are serious (like you couldn't have worked that out LOL). The trim wheel is used to control pitch and the rudders can be used to turn the aircraft (a bit, obviously very carefully, unless you want to induce a spin). You could probably also use aysmmetric thrust a bit too.

 

Abnormal Alternate Law: this mode is nothing to do with failure reversions, it always kicks in if the aeroplane is in an abnormal attitude, so that you can input whatever controls are necessary to effect a recovery to a normal flight attitude, although in theory, if you were in an abnormal flight attitude, it would probably be because Normal Law had dropped out, unless you had maybe spun coming through wind shear or something like that. You could rip the wings off in this mode as well.

 

So, unfortunately as noted by a previous poster, it is entirely possible to stall a FBW Airbus if Normal Law drops out and it reverts to a simpler law, in which case, pulling the stick back will stuff things up just like it does on any other aeroplane. But, even if the pilots on AF447 did not know all that, any pilot worth his or her salt should have thought to try shoving the stick forward when the altitude was dropping faster than a call girl's draws. You learn that pretty much on day one at flying school, so it is quite scary that they apparently were not inclined to try it. As noted on a post above, probably the same blind faith as imagining that the Titanic was 'unsinkable' simply because it could withstand some damage and stay afloat.

 

As I have been moaning about for years to various European authorites, spin/stall recovery training should be mandatory for PPLs, and that should have occurred before you've even got anywhere near an airliner cockpit. Wolfgang Langewiesche wrote his book, Stick and Rudder, over 65 years ago, and in that time, it appears we've lost a bit of common sense airmanship when it comes to training some airline pilots.

 

Al

Amen to that, Al, basic airmanship is a dying skill it seems.

Not to mention that they're flying an aircraft that's supposedly "impossible" to stall - I think there's more than a bit of the Titanic mentality going on when they start saying things like that. I don't think those pilots believed they were really in a stall - there's no other explanation...

I don't think anyone tells such things to Airbus pilots in training. It's the media who latch onto such inaccuracies, just as with the Titanic (never officially described as unsinkable). The FBW only provides stall protection in Normal Law, and AF447 was in Direct Law. The main problem was they didn't recognise the situation they were in, mainly because of conflicting indications and not knowing which indications were reliable and which weren't.

Some things to consider:

 

- Shove the stick forward to break out of the stall might be something you learn on day one of flying school, but basic flying school is the last time you learn it. Airliner stall recovery training (at least pre-AF447) assumed that 1. you would react to the stick shaker or stall warning (i.e. before you were actually stalled) and 2. stalls occur close to the ground (on approach), so the stall recovery technique should focus on the minimum loss of altitude. Airliner pilots are not taught to shove the stick forward, instead they are taught to release back pressure (still far from pull back as hard as you can, but still). Stall recovery was assumed to be such a basic technique that it was not included in any training after the type rating. In the Turkish 737 crash at Schiphol a few years ago it had been something like 18 years since the captain had last practiced stall recovery. Since 'airliners don't get into stalls' it was not a part of the recurrent training syllabus.

 

- Normal law is indeed the only law where you have stall protection. It's also extremely rare for an Airbus to be in anything other than normal law. I remember reading an interview with a pilot of US Airways. He stated that in however many years they've flown A330s they have never had a reversion from normal law in operation. I'm also not sure how often alternate law flying appeared on the semi-annual sim check syllabus.

 

- The stall warning cuts out below 60 KIAS (presumably to prevent nuisance activations on the ground / during the take-off run.) When one of the pilots finally decided to try shoving the nose down (the second co-pilot tried this even before the captain entered the flight deck if I remember the transcript correctly) the plane was already doing less than 60 KIAS, so the stall warning had deactivated. Pushing the nose down caused the speed to climb again and as a result the stall warning reactivated since the plane hadn't broken out of the stall yet.

 

- It's night. You're over a pitch black ocean, possibly experiencing turbulence. Your instruments are not making any sense to you and you know some of them are unreliable (the plane has told you so). You are a human whose brain has evolved to stop you from being eaten by sabre-tooth tigers, not to deal with the fact that the correct way to get out of your plunge towards the ocean is to first push the nose further down. Is it really that strange that you think: I'm going down really really fast! I need to go UP! Go UP darn plane! and yank back on the stick as hard as you can.

Kevin, AF447 was in ALT2 law not Direct law. The whole scenario regarding AF447 is very complex. Yes the PF did screw up but if you look at certain factors you can see that it must have been extremely confusing for them.

 

One example that I brought up in another thread, they could hear a lot of wind noise rushing past the flight deck (they actually commented on this in the cvr) you would not expect to hear that sound at low speed in a stall, the cause of the noise was due to the high vertical speed (10000fpm) but ofcourse they would never have heard that before, and I would imagine 99% of line pilots alive today would never have heard something like that.

 

Increase in wind sound tied to vertical speed is not simulated in the LVL-D boxes and you can test this yourself in good old FS.

 

Put the NGX or whatever into a deep stall, the wind sound will reduce as speed decreases, hold the nose up and listen to what happens as the vertical speed increases.. Nothing happens even though the aircraft could be falling vertically at 100 miles an hour. That is not what it would sound like in the real machine.

 

I'm most certainly not making excuses for the crew however I believe they thought they were in or near an overspeed condition, pitch black, TOGA Thrust, high level of wind noise and unreliable air data computers and PFD indications

I suspect you may be right in suggesting the crew of AF447 may have thought they were in a high speed dive or some such Rob, but one does have to question the idea that if pulling the stick back was apparently not stopping that perceived dive, then it should perhaps cause one to think you might be in a stall, which, if having done plenty of stalls and spins, would make you familiar with the noise phenomenon to at least some degree.

 

However, and again conceding to your observations, as anyone who has ever done any appreciable amount of stall/spin training will confirm, when you get the autorotation off with the rudder and the aircraft is then dropping to regain flying speed, there is indeed a lot of wind nose, since the airspeed builds pretty damn quickly with the nose down that steep, and the automatic reaction then, is to want to pull back fast as the ASI starts winding up really quickly accompanied by that rapidly increasing wind noise and the ground coming up to meet you at a slightly alarming rate. Of course one has to resist that temptation and pull back gently to avoid too many Gs on the pullout.

 

Most of the time when I've done that in real life (which is a lot actually, since I am a strong advocate of its value, even having initiated a spin from a pull up at 1,000 feet AGL on one occasion - don't try that in real life unless you know what you are doing incidentally), I've tended to be gauging the amount of stick pull with the rate the ASI is winding up and the altimeter is winding down, so as to 'hurry up and go slower', and in doing that, I generally like to let the thing get fairly near to its VNE and ease it out, as opposed to bending the wings off the thing and shedding the elevators by pulling up too hard! But again, I've done that in situations where I have a visible horizon when flying small one and two seaters which were stressed for aeros, not airliners, and although I could probably do that sort of recovery without any instruments these days, I wouldn't be quite so confident of pulling that off with an A330 at night with possibly no reliable horizon, artificial or otherwise. Although I do still think that never having done any of that stuff at all leaves a big gap in a pilot's knowledge and arsenal of inherent skills, which is why I'm a bit evangelical about promoting it.

 

Al