Simulator Training for strong crosswind landings

[wangus]

Interesing thread.
I completed my first TR in Jan / Feb (A320) and found the Level D sim impossible to land well, in all scenarios. I got quite stressed out and worried.... I then did base training in 15-20 knots crosswind, 90 degrees to runway, and frankly found it DRAMATICALLY easier to land on centreline, using crab method, then straightening out in flare with rudder followed by sidestick. Sim killed my confidence, the real thing restored it....

[TypeIV]

I'm not nearly as experienced as most of you guys but here's my input regarding the 737.

For applying the correct dose of "crossed rudders", timeing and how the aircraft behaves on the roll-out, I think the simulator is quite good even though a 100% replication isn't possible.

Many people forget that the hardest part of a crosswind landing is making a perfect, sexy, straight and stable roll-out. What difference does a greaser on the center line make if the passengers sitting in the rear part of the fuselage, have their fish and chips sprayed onto the neck rest of the passenger in front of them?

However, hight above the ground, turbulence, aiming points and rotation rate during the flare is much better trained in the actual aircraft, mainly because of the limitations in graphical projection... but hey, if you land the aircraft the way it is written in the FCTM, on these points there are actually almost no differences on how they should be done whether it is a crosswind situation or not. However when the irrational stress level increases, these basic elements of a nice landing seems to suffer.

After a while flying at a homebase, most people get a good feeling on predicting turbulence depending on the season, wind direction vs orography etc, this can of course not be trained in the simulator properly.

[AirRabbit]

Centaurus (– and all the rest of the readers and contributors –)

This is one of the very best threads / series of comments I’ve read in quite a while. The interesting fact is that there isn’t a post in this thread (at least up to this point) with which I would disagree – even slightly! That is truly an amazing statement, from my perspective at least.

Those of us who have been around for a while have witnessed an evolution in simulation from what used to be, to what is available today … but, having been continuously involved in this specific aspect of the pilot training industry, I know, for a fact, that we’ve not reached the limits quite yet. Unfortunately, rather dramatic changes in realism achieved 30 to 40 years ago didn’t take a large amount of technology. Today, huge technology developments, serious amounts of professionally obtained and reduced flight test data, and very imaginative applications of both are regularly combined to achieve barely noticeable advancements in simulations. On the surface, it might appear that such advancements are not as beneficial as they once were. I assure you, this is not so! It’s generally not the easily recognizable changes in fidelity that have the largest impact on a pilot’s ability to recognize and assimilate such changes into his/her performance in either the simulator OR the airplane. It has long-been an understanding in the simulation industry that as the industry advances it takes “more and more” to achieve “less and less” – BUT, the significant issue is that often (not every time … but generally) these “lesser” advances are critical to a pilot’s ability to BOTH consciously AND subconsciously recognize and use the advances in the presentation of the simulated environment. This is because each pilot remains an individual … with individual preferences, and, therefore, individually devised mechanisms that he/she uses to take the information presented in the “real world” into his/her mental processing, from which each individual produces physical reactions/responses, again individually chosen, based on what that individual understands what has historically been used and relied upon to achieve the results that individual has desired. Basically, the goal of a simulator “engineer” is to reproduce in the simulator, ALL of the same things (plural) that happen in the real airplane – in the same sequence and same magnitude. This will allow each individual pilot to use those aspects of the presentations that he/she uses “in the real world” to provide information and allow that individual to make the same recognitions and comparisons he/she uses in the airplane to react/respond in the same manner and sequence he/she would were he/she in the airplane under the same circumstances. Long-time participants on this forum may recall my rather repetitive descriptions of “how to land an airplane.” The goal in those repetitious posts were to demonstrate that it is imperative to allow each individual pilot to pick and use the information sources that make the most sense to themselves – individually. Sure, an instructor can try to teach a student to do whatever task is at hand exactly the way that instructor does it – including the hierarchy of what informative sources that instructor uses (and many do only this). But INDIVIDUALS are not like anyone else – that is the definition of “individual.” My examples of “how to land” were an attempt to show how an instructor should allow the student to pick and choose what information sources are best understood by them – individually – while, at the same time – putting the airplane where it has to be to achieve the best airplane performance. It’s not rocket science – but it does take awareness and understanding. The exact same issues are prevalent in simulation. However, in a simulator, the student will only be allowed to select his/her information sources from those features that are present. If the simulator engineer/manufacturer didn’t include what that student would have selected in the airplane – that student is forced to use what IS available. Now, when that student gets into the airplane, that student is inevitably torn between using what he/she would have preferred to use, and ignoring that preference in order to use what he/she has been taught – which will invariably cause delay and lesser levels of precision and accuracy.

The ultimate is to use simulation that has as much of the real world incorporated into that simulation – such that the simulators response will be as close as humanly possible to the response of an airplane under the same circumstances. That will allow that student to pick and choose the information sources preferred – just as he/she would have chosen in the airplane. This is why I continue to harp on the fact that each instructor MUST know the limitations of the simulator and MUST instruct that student on the correct position in which he/she must place the simulated airplane (just like he/she would do in the airplane) to achieve the desired results.

There are on-going efforts to provide appropriate modeling and data to allow simulators to be able to be programmed in such a way that each student will have the best opportunity to see, and develop the correct response to achieve the correct airplane condition and position for each task … and these areas include gusting winds/crosswinds, aerodynamic stalls, and bounced landing recoveries. Stay tuned for additional simulator advancements!!!

[FullWings]

AirRabbit,

Very interesting to hear what’s happening (or might happen) on the technical sim side.

Do you think we are approaching the simulator equivalent of “uncanny valley”, in that minor (and not so minor) differences between them and reality that used to be all part of the experience are now becoming more obvious because of the overall increase in fidelity? The first sim I “flew” had no visuals, no sound, no motion and clockwork instruments but boy did it seem real at the time, especially to a cadet trying to impress his instructor!

[AirRabbit]

Originally Posted by FullWings

Very interesting to hear what’s happening (or might happen) on the technical sim side.

Do you think we are approaching the simulator equivalent of “uncanny valley”, in that minor (and not so minor) differences between them and reality that used to be all part of the experience are now becoming more obvious because of the overall increase in fidelity? The first sim I “flew” had no visuals, no sound, no motion and clockwork instruments but boy did it seem real at the time, especially to a cadet trying to impress his instructor!

Excellent question, FullWings.

For the unaware, FullWings’ reference to “uncanny valley” comes from the following:
The uncanny valley is a hypothesis in the field of human aesthetics which holds that when human features look and move almost, but not exactly, like natural human beings, it causes a response of revulsion among some human observers. The "valley" refers to the dip in a graph of the comfort level of humans as subjects move toward a healthy, natural human likeness described in a function of a subject's aesthetic acceptability. Examples can be found in the fields of robotics and 3D computer animation, among others. Note the movie, "Avatar."

FullWings … no … what I see has no such implications and, as far as I know, no such implication has been raised by anyone associated with flight simulation at any level. Initially, the motivation for the use of simulation was simple and direct – cost and availability … but the fact was, the simulation used was minimal and its primary contribution was that of “procedural understanding” and not “skill development.” The first sims I flew were just like the ones you describe – “no visuals, no sound, no motion and clockwork instruments” but the fact was there was very little “aerodynamic accuracy.” However, when one considers the benefit to procedural knowledge and understanding, particularly when combined with the existing level of “realism” in the function of the equipment and its relative location, it provided quite adequate practice of cockpit procedures that were notably better than referencing a page out of a manual or even an enlarged photograph mounted on a wall. When the airplane systems “interoperability” in these simulators got to the point of being mostly accurate, another rung on the ladder toward accuracy was achieved.

Actually, with each advancement made (sound, visuals, weather simulation, daylight, the introduction of high speed computers, etc., etc.) another rung (or several) was achieved. However, it was some time before competent aviation professionals actually accepted that “flight training” could be accomplished in a properly designed, constructed, and evaluated simulation device … and even then, there was a requirement that training be “completed” in the relevant airplane, and that the “check ride” would be accomplished in the airplane as well. In the US, anyway, it wasn’t until the very late 1970s that advancements were made in computer science, the methods used in acquiring flight test data, the methods used to reduce that raw flight test data into programs that could be used by the newest, highest speed computers available, and a policy was developed and refined that described the how such flight simulation would be designed, built, evaluated, and used that the US FAA moved very gingerly into the realm of what was called the FAA’s “Advanced Simulation Plan” – consisting of simulation “levels,” only the highest of which allowed full flight training AND the necessary “flight” check, to be accomplished in a flight simulator. Analysis of how do all of this better, more broadly, AND more narrowly in each area has continued over the intervening years.

Sometimes this industry gets the cart before the horse and sometimes it gets focused on one point and let adjacent points slide out of focus – the most recent out-of-focus issue was, described in general terms, as stall recovery. The term was understood differently by major participants, some wanted more details, some wanted “one size fits all,” some were more focused on “procedures to be followed,” some were focused on the accuracy of the simulation. When one considers “the viability” and “the repeatability” of something whose very definition embodies the concepts of “randomness” and “variability” (like an aerodynamic stall) it doesn’t take long to recognize that formulating a computer program to reproduce a “same conditions” profile for those kinds of random and variable events is not something that is easily accomplished – if at all. But very bright minds and very determined folks have been working to be able to provide a simulation that is closer to an actual aerodynamic stall in a simulator than has ever been achieved previously. More of the same is certainly to come – but it’s going to be harder to achieve and once implemented its going to take instructors who are specifically trained on the capabilities and limitations of each simulator they are to use in conducting that training or those checks.

[oopspff7]

AirRabbit.
I'm not a pilot but if I was and I was struggling with a concept you would be the one I would look to for the way forward.

[Tee Emm]

True story from a captain with Korean Air. Forecast wind for landing 15 knots crosswind. Local F/O was PF for the sector so the captain asked the F/O how he intended to conduct the landing. F/O briefed he would leave the automatics engaged until near the flare and then use rudder to align the aircraft with the centre line and lower the into wind wing to prevent drift.

On final it was obvious there was no wind. The autopilot remained engaged as briefed and at the flare the F/O disengaged the autopilot, gave an almighty kick on the briefed rudder and lowered the wing. The aircraft went ape as did the captain, who managed to salvage the messed up landing.

After the captain recovered from the shock of a nearly wiped out landing, he asked the F/O why he had booted in rudder when there was no crosswind? The F/O looked at him amazed and said all he had done was carry out the technique as he had earlier briefed. The captain said but there was no crosswind! The F/O replied it was obvious therefore the forecast was wrong...

[chksix]

Here's a sim platform which could be scaled up to carry an airliner flight deck.
It seems quite capable of providing sideslip etc.

The 401: what is it? - Force Dynamics
Force Dynamics 401CR - Introduction - YouTube

[Denti]

David Learmount posted about a new simulator development a few years ago especially for crosswind training. I have no idea if that concept was mode widely adopted or not though.

Simulators get real - Learmount

[Tee Emm]

Even full flight simulators (FFS), at present, are really just sophisticated procedure and systems trainers, although the aviation authorites pretend they can be beneficial to manual flying skills. But US FAA research says there is no evidence that manual flying skills can be taught in simulators - the skills simply do not transfer to the real aeroplane - See more at: Simulators get real - Learmount

I find that an astonishing statement and hard to believe. I recall the first simulator I flew was during a Boeing 737-200 type rating course in New Zealand 47 years ago. By todays standards it wasn't sophisticated. When on completion of the simulator training I then underwent dual instruction on the real 737, everything came quite naturally to me starting from the first taxiing. In fact the only difference I recall between the simulator and the real thing, was the aircraft I flew had extremely difficult reverse thrust lever movement which was most distracting on the landing run. It was after I flew other 737's where reverse thrust was easily attained that I discovered the first 737 was poorly maintained.

[AirRabbit]

Simulators get real
By David Learmount on 20 July, 2011 in Uncategorised

Cathay Pacific has just become the first customer airline for a revolutionary new simulator motion system that will transform what simulators can be used for.
Even full flight simulators (FFS), at present, are really just sophisticated procedure and systems trainers, although the aviation authorities pretend they can be beneficial to manual flying skills. But US FAA research says there is no evidence that manual flying skills can be taught in simulators - the skills simply do not transfer to the real aeroplane.

Well, my first, off-the-cuff response would be a resounding “HOG WASH!” After taking a breath, I would collect my thoughts and concede that there may be some who would argue (however, I believe, not successfully) that I’ve been living under a rock for the last 45 years – or in some kind of “dream world,” but the fact is that until today, I was completely unaware of Mr. Learmount … and after reading this kind of comment, I’m prone to go back to my “completely unaware” state regarding this particular individual, or at least believe the same comment about his last 45 years. The comment about the “US FAA research” that he referenced is totally unfamiliar to me and because of when, where, and how, I was personally involved with this industry during that same time period, I would immediately categorize that statement as completely false. I say this due to the fact that if there were any such research actually conducted by the US FAA, I am supremely confident that I would have been integrally knowledgeable ABOUT, and very likely involved IN, any, no, every aspect of such research. With this knowledge, I would place this statement right up there with someone claiming that the late Neil Armstrong, the first man-on-the-moon, actually said “…the moon is, indeed, made of cheese.”

Any person knowledgeable about the US FAA’s position with respect to simulation, would be acutely aware of a rather extensive research effort, of some 2 – 3 years duration, under the auspices of something called the “Advanced Simulation Plan,” AND the conclusion reached from that research … which was the publication of 14CFR Part 121, Appendix H, in the summer of 1980, that allowed the use of a specifically qualified Flight Simulator to be used as follows:

Level D Flight Simulator - Training and Checking Permitted: Except for the requirements listed in the next sentence, all pilot flight training and checking required by this part and the certification check requirements of §61.153(h) of this chapter. (my emphasis underlined) The line check required by §121.440, the static airplane requirements of appendix E of this part, and the operating experience requirements of §121.434 must still be performed in the airplane.

We can have diligent discussions about where simulation came from and where it is headed – we can discuss the shortcomings and the positive assistance proper use of simulation actually provides to the successful completion of pilot training … but to state that the skills learned in a simulator do not transfer to the airplane is patently baseless and, dare I say, prejudiced against the use of a very valuable and a very successfully demonstrated training and checking tool. In fact, I am concerned about just how successful the transfer is from simulator to airplane, when some individuals are prone to use what I call “cheat-sheet” numbers or methods that were devised to “get through” a simulator check, and the use of that devised process may not be appropriate when operating the airplane. BUT – and I hasten to add the same caveat that I’ve regularly offered on this forum, and that is that the instructor or evaluator MUST be trained and able to competently and correctly use the specific simulator (including all of that particular simulator’s capabilities and limitations) and that instructor/evaluator remains personally and professionally involved in what is being trained and/or evaluated in that simulator.

[AirRabbit]

Hi oopspff7:

Just a quick note to say "thanks" for the exceptionally kind words!

[the_stranger]

@wangus

I had the exact same issue. Went from an embraer190 to a330 and during my simulator training, I couldn't land the damn thing even if my life would depend on it. Imagine the muscles in my behind being tense when I had to do my first real life landing.

But it was a very descent landing. Writing that off to beginners luck, I was amazed the second one was the same. And so for all the landings I made the last year. Every now and then you get the occasional slightly bad one, but overall they are textbook.

And then I returned to the sim... Horrible again...

There is something wrong or missing in the/that sim. Maybe the seat of the pants feeling, the visual, whatever, but training xwind on a/that sim would be a waste of time for me.

[AirRabbit]

Originally Posted by the_stranger

I had the exact same issue. Went from an embraer190 to a330 and during my simulator training, I couldn't land the damn thing even if my life would depend on it. Imagine the muscles in my behind being tense when I had to do my first real life landing.

But it was a very descent landing. Writing that off to beginners luck, I was amazed the second one was the same. And so for all the landings I made the last year. Every now and then you get the occasional slightly bad one, but overall they are textbook.

And then I returned to the sim... Horrible again...

There is something wrong or missing in the/that sim. Maybe the seat of the pants feeling, the visual, whatever, but training xwind on a/that sim would be a waste of time for me.

Because the simulator is a mechanical device, completely dependent on computer programming, and having all the parts and pieces operating as they were designed, it is certainly possible that something could be wrong or not maintained within the appropriate parameters. However, there is another potentially fallible entity that has to be considered … and that is the operator of the simulator … either the instructor or the person conducting the check. I don’t throw that out to be deliberately controversial or to indicate that instructors or check airmen are not “up to the task.” But, the fact remains, that instructors and check persons are humans just like the pilots – and in order for them to “do the right thing” they have to be trained on all the aspects of the operation of that very sophisticated tool called a flight simulator.

I don’t know who maintains or recurrently examines the equipment you use – but there should be some person or group who are professionally qualified to do just that – and do so on a regular basis. As you would easily understand, there are multitudes of issues that can go wrong or get “out of whack.” Interestingly, the parameters that were selected may not have been the correct ones … maybe continued use has worn parts or interfaces. But equally true is the human link in this process. Instructors (and check persons) have a wide latitude as to how they set up any given scenario and how they may adjust any aspect of that scenario as it unfolds. Most do a superb job of doing precisely this. Others may believe they are doing that superb job and, perhaps through no fault of their own, may have been trained on to how to do something that is either misunderstood, incompletely understood, or, in some remote cases, just plain wrong.

What I do know is that a properly designed, constructed, programmed, maintained, and used airplane flight simulator at Level C or Level D can, and do, offer the pilot the opportunity to be properly and completely trained on crosswind landings – up to, and exceeding, the maximum demonstrated crosswind limits of the airplane. That’s not just speculation – and that is not just my opinion. I have personally flown more airplane simulators than most of my colleagues – from the very basic visual simulators (in fact, long ago I flew non-visual and even non-motion simulators) all the way up to the most modern and most advanced Level D simulators. The older the simulator, the lower qualification level, usually presents more “challenges” in doing the most “at-the-edge” kinds of flight performance … but there isn’t (or their shouldn’t be…) any Level C or Level D simulator out there in which you wouldn’t be able to fly and land in the most demanding crosswind condition authorized for that airplane and it would be (should be…) as close to your expectations as you would like it to be. Even the oldest airplane flight simulator in existence in the US – a Level A simulator for the “Jetstar” airplane, can be used to instruct and practice crosswind landings. This is not just a hopeful circumstance that only some simulators happen to meet … I’ve always operated on the premise that if the simulator doesn’t do what it is supposed to do (and you have to understand that the lower the level of simulation … the lower the level of fidelity … and, potentially, accuracy – all of which has to be understood and taken into consideration by the organization and primarily and specifically, the instructor/check person using that specific device) then that simulator should be taken out of service (or at least restricted with regards to what it can be used for) while it is examined, repaired, reprogrammed, or whatever other detail has to be accomplished or re-accomplished, before it is returned to full service authorizations.

In case you were wondering - I am a HUGE advocate of properly constructed simulation to train and test pilots, and provide for each pilot, the confidence that they need to have to do the job we all expect them to do. We cannot accept lackluster simulation. We cannot accept improper or incomplete dependence on simulation to do the teaching. A simulator is a tool - a very special tool, no doubt, but a tool nonetheless. And it should be a tool that can be relied upon to do what we expect it to do - or we need to stop using it until it can be brought up to standards.

[Judd]

I am a HUGE advocate of properly constructed simulation to train and test pilots, and provide for each pilot, the confidence that they need to have to do the job we all expect them to do.

During in-flight (not simulator) training on the 737-200 a long time ago, we were given simulated engine failures after V1 by closing one thrust lever to idle and completing a circuit and asymmetric landing or a go-around. The FD was the Collins FD 108 which we switched off for the purpose of the exercise. The engine failure climb out was quite easy to handle.

I found this in stark contrast to later versions of the 737 (Classics) where the aircraft was equipped with the now almost universal "twin needles" FD. Despite many hours in 737 Classics simulators, I have always experienced great difficulty trying to "fly" the FD needles at instant of engine failure and initial climb to flap retract height. Although well aware of the folly of `chasing` the FD needles during engine failure climb out, I find it well nigh impossible to "look behind" the FD at the aircraft attitude. In fact I do much better if I glance over and fly the miniature standby ADI, which is unencumbered by needles, to maintain an acceptable climb out attitude.

While this thread is about crosswinds operation, it has also developed into most interesting discussions on simulator fidelity in general. Hence my comment on FD use in initial one engine climb. Could this be a simulator FD fidelity issue? Are the needles too sensitive to the slightest yaw or roll thus causing over-reaction by the pilot? If I can do a perfectly accurate raw data engine failure at V1 and initial climb out without over-controlling, why is it that with FD operating I am all over the sky trying to pin the needles to the centre? Frankly, I must admit the FD spooks me with engine failure and I suspect a lot of other pilots too. In my book the FD is designed to be an aid - not a hindrance.

[Bergerie1]

AirRabbit


Thank you for your excellent posts, we rarely get ones of such quality.


I spent many years as an IRE/TRE on the three aircraft types I mentioned at the start of this thread. I was also lucky enough to fly with test-pilots to evaluate simulators and trying to get one in particular, in the 1970s, to fly like the real aircraft after training pilots had tried to improve it - but hadn't!


Your knowledge of level D simulators is vastly superior to mine but I share your confidence in their ability to provide good training for cross-wind landings. However, training at or after the stall is a different matter and I would very much like to hear your views on the subject.

[misd-agin]

The planes fly better than the simulators do(17,000 hrs on 10 different models).

Modern Level D simulators do an excellent job of normal flying. And can be used for teaching manual flying skills. But they'll never 100% replicate flying the real airplane. A simple fact is simulators don't scare you but airplanes can kill you.

But the simulators are good enough that the first flight for newly qualified pilots is normally with a full load of passengers.

Simulators built in the last couple of years are better than 3, 5, or 10 years earlier. Flying an old level D simulator is horrible compared to a new build simulator. Is the old model good enough? Yes. Is it at the level of a new build simulator? Absolutely not.

[the_stranger]

@airrabbit

While I am not exactly sure who maintains our simulator, it is done inhouse. I fly for a large (west) European airline and know from other (non simulator) processes, the "government" is really demanding a lot of procedures, documents and checking/testing, so I assume, but again, do not know for sure, the simulators are also routinely checked and properly maintained/programmed.

As far as operating the thing, I only "fly" it once every few months, do I can't comment on it, but assuming it is working properly, how can you screw it up as an operator when you input 25kts xwind? Again, I have no experience "pushing" the buttons, but I do have experience in the fact our simulators (for this type of aircraft) give me a lot of bad landings, while in real life my landings go fine.

Whatever the reasons, I can't properly land the simulator version of my aircraft. And I am not alone. When there was too little in-house capacity to train pilots, they trained them at other airlines in other countries. Different simulator, different operators, different regulators, but same experience (for some at least). Posts in this topic tell me the same, xwind training on simulators do not always work, be it the simulator, the maintaining, the operating, the pilots themselves, but the only thing my simulator training taught me was to "fear" xwind. Fortunately, real life taught me something else.
(I am specifically talking about landings, not other aspects of training. Besides that, the training does give you a rough outline of the technique used on landing. But was I prepared, sufficiently trained and confident to do a 25kts xwind on my first flights? No, not even close. After a year of flying, I get closer to that point, but purely based on experience gained in real life, not simulators).

[de facto]

After a year of flying, I get closer to that point, but purely based on experience gained in real life, not simulators).

Probably because you spent more time in the aircraft than you did in the sim?

The simulator allows you work on your scan,and set up your method how to deal with a crosswind,possibly a decreasing crosswind as you go down,hence a possible quick change in track required to maintain the centerline..the requirement to focus inside and outside....
The feel of the pants i believe is only a benefitial aid during low visibility,poor lighted runways manual landings...the rest is all about scan,looking outside including for the flare (pitch up),rudder input to straight up the nose and sink rate the rate at which one should decrease the thrust...
A croswind requires in general a longer landing distance as the pilot decrabs and the thrust stays in a tad longer...

[A37575]

Posts in this topic tell me the same, xwind training on simulators do not always work, be it the simulator, the maintaining, the operating, the pilots themselves, but the only thing my simulator training taught me was to "fear" xwind

While the various posts indicate some concern about the validity (fidelity) of simulators in crosswind landings, maybe they are better than the alternative ie nothing. Can you just imagine the OMG reactions of the captive audience down the back when a MPL cadet pilot is given his first 30 knot crosswind landing in the real aircraft without the benefit of previous simulator training to get the technique right? He may be nothing more than the captain's apprentice yet he is legally second in command.

Of course all this is hypothetical but all I can say is thank goodness for todays sophisticated simulators.