Simulator Training for strong crosswind landings

[AirRabbit]

Originally Posted by rudderrudderrat

You are not alone. We must use our sense of lateral acceleration (poor in most sims) and heading changes (non existent in most sims) to anticipate track deviations earlier. When those clues are missing or lag the real world, then we tend to over correct which causes PIO.

& I'm glad this effect is now admitted:
"Many pilots also develop nausea in simulators due to the poor correlation between visual and motion cues."

Of course it is true – and has been acknowledged, for well over 3 decades, that visual cues that precede or lag an associated motion on-set cue can, and does in many cases, cause some degree of discomfort – and that often varies widely from person to person. However, there is more than a modicum of a requirement to accept that what is seen and what is felt is, indeed, mutually generated causes-effects … AND … because the visual systems are adjusted to provide the appropriate cueing to pilots who are occupying the pilot seats, and those seats are properly adjusted to the appropriate pilot eye-position, those who see the visual scene from a position other than the pilot’s position, have a greater potential for suffering from this effect.

I suspect that you are very likely aware that as a requirement of simulator evaluation for qualification – by all of the regulatory authorities with whom I am familiar – all require that the motion and visual systems fall within a published range of initiation/response. Also, there are times when, for whatever reason, this authorized range of initiation/response has become compromised – and the required daily preflight would likely catch this anomaly, but surely if such a distracting situation were to be written up in the simulator discrepancy log – the maintenance folks are usually quite good at reviewing and correcting those kinds of issues.

While I acknowledge that some simulators don’t provide lateral cues that are as good as the same cues in an airplane, it is also true that they do exist, in each and every simulator – perhaps not to your liking or expectation – and, to some degree, the LM2 addition can provide a better lateral cue – but, as I described in the above post, it does so with at least the potential of compromising the roll cue, and that potential is either not realized, or is realized, or is not only realized but is exacerbated, depending on the individual experiencing the phenomenon and to what degree the roll cue has been compromised by that programming.

[Miserlou]

With the level of technical knowledge here, I wonder if I may ask if anyone has any data about friction on crosswind landing and take off, more specifically, the difference between the nose and main wheels.

We have been having an ongoing discussion about whether the main wheels would give up first allowing the aircraft to be blown off the downwind side of the runway or whether the nosewheel would give up first allowing the aircraft to weathercock toward the upwind side of the runway.

It would appear from the copious video evidence on youtube that there are several variations but seems to boil down to two scenarios.
1. The upwind wing is allowed to lift (poor aileron technique usually) which allows the whole aircraft to be pushed toward the downwind side.
2. The into wind wing remains level or low but poor rudder technique or excessive gust strength cause the nose weathercock.

Scenario 2 seems the most common. Any other views?

[rudderrudderrat]

Hi AirRabbit,

Thanks for your explanation for the reluctant introduction of LM2.

You say “Note: a typical person will feel the cab rolling when it is rolling at 2 degrees/second or more.”

Similarly, I would have thought a typical pilot could also sense the cab pitching at 2 degs/second or more. During take off, the cab rotates backwards quite steeply and quickly to simulate the acceleration on the runway. (The converse is true during braking.) With strong visual clues, we pilots interpret that as an acceleration and not as a rotation.

Similarly prolonged sideways accelerations can be simulated with cab roll and provided we have strong visual clues, we would interpret the roll as a sideways acceleration.

One of your reasons for not accepting LM2 is “However, this is achieved by limiting the roll capability of the simulator, resulting in a situation where the simulator cab cannot roll much to represent the aircraft's roll, effectively decreasing roll fidelity”

I was not aware that the cab rolled at all during a balanced banked turn. Please could you expand some more.

[piratepete]

AIRPLANE FLYING vrs SIMULATOR TECHNIQUE

Its really quite simple.In an airplane we absorb all the cues that come our way and respond to them.

In a simulator, what the machine does is directly related to input ("fidelity").Once you have memorised the NUMBERS, PITCH ATTITUDE, POWER SETTINGS, CONFIGURATION etc etc, you will always get the same result.Simulator "technique" will never be the same as real world technique, this is not possible.Real world is too complicated and dynamic.

[AirRabbit]

Originally Posted by rudderrudderrat

Thanks for your explanation for the reluctant introduction of LM2.

You say “Note: a typical person will feel the cab rolling when it is rolling at 2 degrees/second or more.”

Similarly, I would have thought a typical pilot could also sense the cab pitching at 2 degs/second or more. During take off, the cab rotates backwards quite steeply and quickly to simulate the acceleration on the runway. (The converse is true during braking.) With strong visual clues, we pilots interpret that as an acceleration and not as a rotation.

Actually, the acceleration and rotation for takeoff cueing typically works as follows (and there are some alterations, substitutions, and variances from manufacturer to manufacturer and operator to operator): the initial acceleration of the airplane is replicated closely by the platform moving forward and the movement is at such a rate that it is recognized by the on-board crew. However, the platform jacks cannot go on moving without reaching their “limit stops” and a technique is used that prevents the stops from being reached without being discernible to the crew. This technique is to gradually decrease the displacement rate (known as a “washout”) where that rate of change is below the sensory threshold of the simulator occupants (typically called the “recognition threshold”), eventually reaching the zero, and begins “resetting” the simulator to the “neutral” position fore/aft (again, below the recognition threshold rate).

Sometimes, a simulator may be adjusted in "pitch" - again, below that threshold rate, to provide a cue that the simulator is continuing to accelerate. Depending on the rotation angle for the takeoff/climb, the rotational angle of the simulator cab, may or may not be gradually lowered. If the cab remains tilted up, it will provide a sensation of continuing to accelerate, but if the attitude is not significant, the acceleration rate will not be “felt” as significant either, although it will be present and very likely noticable. At the point where the simulated airplane is to be rotated to achieve a climb angle, the simulator cab is rotated, this time at a rate that is noticed by the on-board crew, and the rate of rotation is gradually reduced, again below the threshold rate. At some logical point during the first portion of the climb, the rotation angle is eventually reduced all the way to zero – removing all acceleration and rotation cues.

Originally Posted by rudderrudderrat

Similarly prolonged sideways accelerations can be simulated with cab roll and provided we have strong visual clues, we would interpret the roll as a sideways acceleration.

Unless you are describing something unusual, I think you might want to re-examine that conclusion. In an airplane, a sideways displacement – either to the right or to the left, is normally not a long-lived experience. Flat, long term, level movements are not normally experienced, and when they are, they are almost always short lived, brief duration excursions, usually caused by disrupted airflows, and are typically recognized in the airplane as little more than “turbulence,” and turbulence models are typically available in a simulator.

If you’re describing a level turn, either right or left, in the airplane, the occupants would physically recognize the left or right bank, and from that point forward the prominent force recognized is more of an increase in “g-forces” to maintain a constant altitude and is the reason your coffee doesn’t spill in your lap. To have the simulator accurately replicate an airplane level turn, either left or right, you would want the simulator occupants to recognize the on-set cue of the initiation of the airplane bank, but as the desired bank angle is approached, the bank angle should be removed, and return the simulator cab back to a level attitude ... all accomplished below the recognition threshold ... at the same time an increase in simulator nose-up attitude is also achieved ... also below that same threshold. By doing this the force of gravity would provide a recognition by the simulator occupants as an increased “g-load,” as would be recognized in an airplane under the same circumstances.

Rolling out of that turn would essentially be achieved in a simulator by reversing the sequence of the adjustments used to achieve the simulation of the turn in the first place; i.e., reduce the pitch attitude of the simulator (again, below threshold), provide a simulator cab roll in the opposite direction (simulating a roll-out of the bank angle), and again after initiation of the on-set cue, remove that input, again below recognition.

Originally Posted by rudderrudderrat

One of your reasons for not accepting LM2 is “However, this is achieved by limiting the roll capability of the simulator, resulting in a situation where the simulator cab cannot roll much to represent the aircraft's roll, effectively decreasing roll fidelity”

Please understand, it’s not that I don’t “accept” LM2 – it’s just that the technical folks with whom I’ve worked regularly, tell me that today's existing algorithms can provide a better mix of roll and lateral cueing than LM2 can provide … and doing this would achieve the same, or better, lateral cueing as provided by LM2, and do so without compromising the rolling cues that are present all the time with existing programming; where by using LM2, the roll rate cueing is likely to be compromised in all situations where a roll is accomplished. However, the fact is that one has to know what they are doing to achieve that “better” mix, and most people do not have the experience in doing that. It is for these reasons that the FAA allowed LM2 to be used, but refused to mandate its use.

Originally Posted by rudderrudderrat

I was not aware that the cab rolled at all during a balanced banked turn. Please could you expand some more.

Each time the flight crew initiates a bank angle, the simulator is actually rotated (banked) in the direction of the airplane’s bank (and very likely also tilted up, but at a rate below threshold) to indicate a slight increase in “g-force,” with the simulator cab rate of roll and bank angle being washed out, again, below the recognition threshold, and returned to neutral. When the simulated airplane is rolled to a wings level attitude, the simulator cab is rotated in the opposite direction, to provide that rolling on-set cue, and then both the roll rate and roll angle is washed out and the position is reset to neutral – both, again, below the recognition threshold … as I indicated above.

[rudderrudderrat]

Hi AirRabbit,

Thanks very much for the info on sim motion cueing.

Flat, long term, level movements are not normally experienced, and when they are, they are almost always short lived, brief duration excursions,

During taxiing with big turns, we experience long term sideways accelerations. During the take off and especially during a Rejected Take Off, the sideways accelerations in the sim are not realistic and often result in over controlling and PIO. (Please see article by Learmount)

it’s just that the technical folks with whom I’ve worked regularly, tell me that today's existing algorithms can provide a better mix of roll and lateral cueing than LM2 can provide …

Well I for one am really looking forward to your technical folks providing better algorithms. How long do you think we will have to wait?

[AirRabbit]

Originally Posted by rudderrudderrat

Thanks very much for the info on sim motion cueing.

You are very welcome.

Originally Posted by rudderrudderrat

Originally Posted by AirRabbit

Flat, long term, level movements are not normally experienced, and when they are, they are almost always short lived, brief duration excursions,

During taxiing with big turns, we experience long term sideways accelerations. During the take off and especially during a Rejected Take Off, the sideways accelerations in the sim are not realistic and often result in over controlling and PIO. (Please see article by Learmount)

Well, first, I apologize … I wasn’t aware you were speaking about ground based operations. However, I can say that I don’t believe I’ve encountered the kinds of PIO you describe (i.e., during Take Off or Rejected Take Off) in any simulator. So, I’m somewhat at a loss to know exactly what you mean.

Second, I am familiar with David Learmount, and his article on LM2. I’m also familiar with Capt. Filip VanBievliet, the guy who “developed” LM2 (Lateral Maneuvering Motion), while he was with Sabena – and it was Sabena that was originally offering LM2 for sale. Mr. VanBivliet, an engineer by training, has, I think, spun off his company from Sabena, and I, believe, is now the owner/principle operator of “AWx” or “Acceleration Worx” – still out of the Netherlands. While I have no personal issues with Mr. VanBivliet and have no intent or desire to insult or criticize him or his work, I do know some in the business who would take that opportunity. Of course, I’m sure I have my own detractors out there as well. If you would like to know a bit more about what at least some in this industry have to say on this subject – please send me a PM and I’ll share some of them with you. I understand if this would not be of interest to you.

Originally Posted by rudderrudderrat

Originally Posted by AirRabbit

it’s just that the technical folks with whom I’ve worked regularly, tell me that today's existing algorithms can provide a better mix of roll and lateral cueing than LM2 can provide …

Well I for one am really looking forward to your technical folks providing better algorithms. How long do you think we will have to wait?

I suppose, as you would likely imagine, that would depend on the level of urgency those folks feel about developing such algorithms … and what it would mean to them … and that likely would be directly related to either job security or some other financial benefits. Of course, a regulatory authority could mandate that such improvements be made … and the likelihood of that happening is also directly tied to the amount of public clamour that might be generated. So, I don’t have an answer for you … but I suspect it is likely to happen – sometime.

[rudderrudderrat]

Hi AirRabbit,

However, I can say that I don’t believe I’ve encountered the kinds of PIO you describe (i.e., during Take Off or Rejected Take Off) in any simulator.

I have quickly scanned this page and the first two pages of this thread and found these comments:

Jwscud #104
"The ground handling was very over-sensitive, with PIO being very easy to induce on the takeoff or landing rolls."

FlightGlobal "Lm2 is a patented software solution that modifies the conventional lateral accelerations applied by six-axis motion systems in full flight simulators – which frequently cause trainees to overreact, resulting in pilot-induced oscillations."

safetypee #12
"In addition to the many weaknesses in simulating crosswinds as above, the majority of systems are unable to simulate true lateral acceleration, at least for a significant period covering de-crabbing and during roll out. ... because there is no sideways ‘seat of the pants’ feeling for feedback."

wangus #22
"I completed my first TR in Jan / Feb (A320) and found the Level D sim impossible to land well, in all scenarios.... Sim killed my confidence, the real thing restored it"

Many experienced pilots (I would say over 50%) find the ground handling of the sim during take off & landing far more difficult than real life. It sounds like you require the regulatory authority to mandate that such improvements be made before it will happen.

I think we need more competition between the sim manufacturers - like there is between the airlines.

[AirRabbit]

Originally Posted by rudderrudderrat

Originally Posted by AirRabbit

However, I can say that I don’t believe I’ve encountered the kinds of PIO you describe (i.e., during Take Off or Rejected Take Off) in any simulator.

I have quickly scanned this page and the first two pages of this thread and found these comments:

Jwscud #104
"The ground handling was very over-sensitive, with PIO being very easy to induce on the takeoff or landing rolls."

FlightGlobal
"Lm2 is a patented software solution that modifies the conventional lateral accelerations applied by six-axis motion systems in full flight simulators – which frequently cause trainees to overreact, resulting in pilot-induced oscillations."

safetypee #12
… are unable to simulate true lateral acceleration, at least for a significant period covering de-crabbing and during roll out. ... because there is no sideways ‘seat of the pants’ feeling for feedback."

wangus #22
"I completed my first TR in Jan / Feb (A320) and found the Level D sim impossible to land well, in all scenarios.... Sim killed my confidence, the real thing restored it"

Many experienced pilots (I would say over 50%) find the ground handling of the sim during take off & landing far more difficult than real life. It sounds like you require the regulatory authority to mandate that such improvements be made before it will happen.

First, the comment you quoted from FlightGlobal sounds very much like the sales pitch offered by AWx. Again, it is not my intent to impugn the reputation of FlightGlobal or that of David Learmount, but that quote sounded very much like it could have been part of an advertising brochure. Additionally, as I indicated previously, I have it on pretty good authority that those specific words have been used by the developer of this algorithm, and his “forceful” personality is apparently approaching legendary status in the world of simulation engineering “geeks.” So, an article that was specifically looking at this particular product, using this quote (sounding like something that the company representatives say, in that the algorithm apparently does pretty much what it is advertised to do) is not at all surprising. However, like I indicated earlier, of those in the industry who acknowledge what this algorithm does, there are some who point out characteristics that can point to weaknesses or compromising aspects of the same program.

Second, notwithstanding the quotes from Jwscud, safetypee, and wangus, and without meaning to criticize or disparage any of them or what they say, and while I fully acknowledge that there are specifics of almost any simulator about which some pilots can, and often do, offer criticisms (some of them very specific, some quite general, but all of which certainly convey personal concerns or objections), and without attempting to sound self-centered or anything approaching some kind of “paragon of truth or accuracy,” the fact is that I have had the opportunity to fly a substantial number of simulators, not only in the US but around the world, and most of the airplanes they represent. Certainly, this does not come close to including every simulator currently in service. However, as I said, the simulators I have flown have not demonstrated the kinds of PIO during Take Off or Rejected Take Off described in these pages. Now, that may be because I have flown such a broad spectrum of so many airplanes and simulators, that I’ve become immune to some of the minor attributes that often crop up in any given piece of machinery built to replicate/simulate another piece of machinery; however, I’m not at all convinced that this is the case.

What I am saying is that if any regulatory authority were to believe that the competency of pilots being trained and/or checked using simulators that have the kinds of problems described here, is being compromised, they would likely have little option but to either remove the authority to use simulation, or those specific simulators, or at least those specific tasks from those specific simulators for the training and/or checking of those pilots.

To make the case that has been claimed here, someone would likely have to document it ALL, as completely as possible, and as clearly as possible, and present that information to whatever regulatory authority is appropriate. If that regulator were to believe that a sufficiently robust case has been made, they would likely share that with the other regulatory authorities with whom they regularly confer. Again, presuming the accuracy of the data and the potential issues that might arise as a result, it would make sense that the regulatory authorities would, en masse, mandate that such lateral cueing be improved. Again, it is my opinion that no regulatory authority would require any simulator sponsor to purchase a particular product from a particular source – rather they would invariably address the technical requirements that would have to be met in order to achieve the level of fidelity deemed appropriate, and leave each sponsor to their own methods of achieving those ends.

The other alternative would be to make the same case to individual airlines or simulator sponsors (either one at a time – or several of them together) and convince them that additional fidelity is required (in this case, lateral fidelity) to better ensure that crewmembers would be trained and tested against appropriate standards using equipment that allows those standards to be met using the newly established fidelity standards. The question that would wind up being “the elephant in the room,” is why should they take this position and spend whatever monies might be required to either purchase the programming and appropriate support, or provide for the additional training of their own simulator maintenance/engineering staff so that their own personnel would have the capabity to provide that same end. The remaining issue would be cost – since actually doing this would be completely voluntary. The answer to this issue is very likely what I was told a hundred years ago by a very astute trainer … and that is “while training is expensive, it is nowhere as expensive as an accident.”

Beyond those alternatives, I have only one additional relevant suggestion ... below.

Originally Posted by rudderrudderrat

I think we need more competition between the sim manufacturers - like there is between the airlines.

The competition factor between simulator manufacturers is probably one of the very few logical events, other than the end points I’ve pointed out, above (either regulatory requirement or sponsor recognition of competency to avoid accidents) that might achieve the end points that have been outlined in your posts.

If the simulator manufacturers were to be put on notice that further simulator acquisitions would be directly dependent on any such purchases including the established level of lateral fidelity (and you would very likely be asked to provide some specifics with respect to what fidelity measurement you would find acceptable) the degree to which this might be realized is the degree to which simulator manufacturers were factually convinced that without such additional fidelity their simulator sales would likely suffer. Logically, such suffering would be in direct relation to the level of support that could be mustered from current and future simulator users/purchasers or those who use the services of simulator manufacturers for their maintenance and servicing requirements. The caveat is, clearly, that without the simulator manufacturers being able to independently verify the accuracy of customer demands for this additional fidelity, such ... well, threats ... would very likely be essentially worthless.

[Linktrained]

Newbies used to gain some INSTRUMENT FLYING practice on a Link Trainer. This did not have an A/P, so one learned to fly manually, using the somewhat basic instruments in front of one. These could be further restricted to just a Limited Panel.
Earlier exercises had one flying a list of timings, speeds and headings to produce a drawing of a hand on the Instructor's table.
Instrument let downs with QDM or ADF patterns could be practiced with various strength of W/Vs, down to a low approach level, followed by the overshoot procedure, as one did not have the R/W in sight... The hood was still in position !
The equipment could be set for " Calm" or a rudimentary " Rough Air".
Many Newbies stalled and spun the trainer and learned that recovery on instruments was possible ( even if not quite perfect).

[Jwscud]

Airrabbit, I understand your points completely.

However, as a bloke of limited experience who has only sat in the right seat for a shortish period of time and only have two type ratings, in some respects I am more the target demographic for Level D sims than yourself. A simulator that is slightly unrealistic and over-sensitive to the point of being able to induce PIO in a pilot without the benefit of your experience, or on a new size or type of aircraft is manifestly not doing its job properly.

One is able to achieve a Type rating solely through training in the sim, then 45 minutes flying the aircraft to do the required circuits. Therefore the significance of any handling defects is amplified. Obviously in this case it makes life easier as the sim is hard to handle but the aircraft significantly easier. It was however a source of unneeded frustration and distraction on the type rating. It also wastes time on Line Training, as with no experience on type, one is not well positioned to differentiate between solely sim flying techniques and how to actually fly the aircraft as you have advocated previously.

Call me a cynic, but it is also not in the interest of sim manufacturers to document flaws in approved sims in a way that might require investment of large quantities of cash to fix their flight models slightly. Equally, minor issues (I agree that these don't particularly affect the sim overall) may indeed be masked by the skill of the pilots testing the simulators.

[Linktrained]

P.S. to #116,
Drawing the harid took about an hour and any lack of precision would show up.
On the outbound part of the let down one tried to assess the drift, perhaps 10 degrees Port, which would be something like 10 degrees Starboard on the final approach. (" Try and remember !")
The Magnetic Compass would have shown acceleration errors, as would the ball or slip needle ( these may well have been artificially induced). Flying by the seat of one's pants was NOT the right thing to do.


By chance I went to the French National Gliding Centre, where landing along the wide top of a ridge, which usually had a cross wind ( useful for ridge soaring for endurance flights) The cross wind for landing must have been more than half one's landing airspeed, landing on the one wheel or skid and hoping to lower the into wind wing in the last moments of the landing run. ( The instruction was in Franglais, both learning !)
Some of the landings were up the slope and with the usual ( now) tail wind, aiming to stop close to the top for a turn around for a further flight. Interesting, Not airline!

[AirRabbit]

Hi Jwscud

Thanks for your understanding – and, let me return the compliment – you make your points both clearly and succinctly – and, having done so, it is evident that there is being generated at least some (perhaps a great deal MORE than “some”) confusion and questions. I don’t know of anyone who claims to be a professional aviator who would accept the kinds of questions and the degree of confusion you, and the others on this thread, have described.

Not that it will make any difference to anyone, yourself included, but I can acknowledge that the level of concern expressed over these last few days (weeks? more?) has not gone unnoticed, at least by me. Of course, I certainly cannot even suggest that one little ‘ole voice (mine) could do much of anything toward addressing those question or those concerns, as the “bully pulpit” I now occupy has nowhere near the “bully-ability” as the one I used to occupy … but, let me say this … I AM going to discuss these issues with colleagues of mine – some of whom are still integrally involved in this particular business – and the others, in addition to having their own theories and opinions, also have their own colleagues and professional contacts.

I have been impressed with what has been said – particularly if I filter out the rhetoric and the emotions (both of which are totally understandable, by the way), and I am now pondering …
1) what I may be able to do to advance the concerns you have expressed and do so with those who might be in positions to do something meaningful about it … and
2) just how I might be able to present that information in a manner that would allow a better grasp of the facts involved.

I would love to be able to guarantee that whatever it is I am able to generate will, indeed, “fix the problems” you’ve outlined. Unfortunately, all of whatever I can generate may only be acknowledged and dismissed – although, I think the dismissal option is really quite minimal. The question might be, not only what can be done, but might be the cost involved and who, logically and correctly, should be required to bear that cost. As I’m quite sure you know, as we say here, in the Colonies, “there is no such thing as a free lunch.” But check back once in a while and I’ll try to keep you abreast of whatever is going on.

Best regards,

A/R

[safetypee]

To add to the deficulties of simulating crosswind landings, consider ground effect. At best (still wind), good quality in-flight data is difficult to obtain; it would be interesting to know what data (or assumptions) is used in simulating ground effect with the effects of drift or turbulence.

A further thought, in real conditions most pilots’ experience of crosswinds is gained by extending their previous limit, this differs from most other operations where conditions are often repeated and, perhaps excepting turbulence, are similar. For crosswinds, some conditions can be a significant extension of previous landing experience (#1), thus with additional demands in turbulence, perhaps there is some apprehension (the unknown) driving the need for quality simulation.

[AirRabbit]

Originally Posted by safetypee

To add to the deficulties of simulating crosswind landings, consider ground effect. At best (still wind), good quality in-flight data is difficult to obtain; it would be interesting to know what data (or assumptions) is used in simulating ground effect with the effects of drift or turbulence.

A further thought, in real conditions most pilots’ experience of crosswinds is gained by extending their previous limit, this differs from most other operations where conditions are often repeated and, perhaps excepting turbulence, are similar. For crosswinds, some conditions can be a significant extension of previous landing experience (#1), thus with additional demands in turbulence, perhaps there is some apprehension (the unknown) driving the need for quality simulation.

Hi safetypee
Obviously, I am unaware of how familiar you might be with the simulation requirements of your, or any other, regulatory authority, so ... here, for your reading pleasure (or for treatment of your insomnia) is the US FAA requirements – initially contained within Advisory Curculars – until their legal officers let them know that they could not withhold the issuance of a qualification or interfere with the issuance of an approval to use a simulator simply because it didn’t match a “recommendation,” which is all that an Advisory Circular really is. That is the reason the FAA developed and published these requirements as a full-fledged regulation, 14CFR Part 60.

For information, here are just a few of the relevant requirements (at least now they're requirements) regarding “ground effect” and “crosswinds” as found in the above-referenced rule. Suffice it to say that the amount of data and the method of collection has been through several trials, errors, corrections, trials again, modifications, etc., etc. The bottom line is that what is produced today, and essentially for the last decade or two, while it may have some inevitable shortcomings, is actually pretty darn good data.

From the FAA Part 60 Regulation:

OBJECTIVE TESTING FOR CROSSWIND TAKEOFF.

TOLERANCE: ±3 kts airspeed, ±1.5° pitch angle, ±1.5° angle of attack, ±20 ft (6 m) height, ±2° bank angle, ±2° sideslip angle; ±3° heading angle. Correct trend at groundspeeds below 40 kts. for rudder/pedal and heading. Additionally, for those simulators of airplanes with reversible flight control systems: ±10% or ±5 lb (2.2 daN) stick/column force, ±10% or ±3 lb (1.3 daN) wheel force, ±10% or ±5 lb (2.2 daN) rudder pedal force
FLIGHT CONDITIONS: Takeoff
TEST DETAILS: Record takeoff profile from brake release to at least 200 ft (61 m) AGL. Requires test data, including information on wind profile for a crosswind (expressed as direct head-wind and direct cross-wind components) of at least 60% of the maximum wind measured at 33 ft (10 m) above the runway.
INFORMATION: In those situations where a maximum crosswind or a maximum demonstrated crosswind is not known, contact the NSPM.

OBJECTIVE TESTING FOR CROSSWIND LANDINGS.
TOLERANCE: ±3 kt airspeed, ±1.5° pitch angle, ±1.5° angle of attack, ±10% or ±10 ft (3 m) height ±2° bank angle, ±2° sideslip angle ±3° heading angle. Additionally, for those simulators of airplanes with reversible flight control systems: ±10% or ±3 lb (1.3 daN) wheel force ±10% or ±5 lb (2.2 daN) rudder pedal force
FLIGHT CONDITIONS: Landing
TEST DETAILS: Record results from a minimum of 200 ft (61 m) AGL, through nosewheel touch-down, to 50% decrease in main landing gear touchdown speed. Test data must include information on wind profile, for a crosswind (expressed as direct head-wind and direct cross-wind components) of 60% of the maximum wind measured at 33 ft (10 m) above the runway.
INFORMATION: In those situations where a maximum crosswind or a maximum demonstrated crosswind is not known, contact the NSPM.

OBJECTIVE TESTING FOR GROUND EFFECT.
TOLERANCE: ±1° elevator ±0.5° stabilizer angle, ±5% net thrust or equivalent, ±1° angle of attack, ±10% height or ±5 ft (1.5 m), ±3 kt airspeed, ±1° pitch angle.
FLIGHT CONDITIONS: Landing.
TEST DETAILS: The Ground Effect model must be validated by the test selected and a rationale must be provided for selecting the particular test.

REQUIREMENTS FOR OBTAINING GROUND EFFECT DATA:
a. For an FFS to be used for take-off and landing (not applicable to Level A simulators in that the landing maneuver may not be credited in a Level A simulator) it should reproduce the aerodynamic changes that occur in ground effect. The parameters chosen for FFS validation should indicate these changes.
(1) A dedicated test should be provided that will validate the aerodynamic ground effect characteristics.
(2) The organization performing the flight tests may select appropriate test methods and procedures to validate ground effect. However, the flight tests should be performed with enough duration near the ground to sufficiently validate the ground-effect model.
b. The NSPM will consider the merits of testing methods based on reliability and consistency. Acceptable methods of validating ground effect are described below. If other methods are proposed, rationale should be provided to conclude that the tests performed validate the ground-effect model. A sponsor using the methods described below to comply with the QPS requirements should perform the tests as follows:
(1) Level fly-bys. The level fly-bys should be conducted at a minimum of three altitudes within the ground effect, including one at no more than 10% of the wingspan above the ground, one each at approximately 30% and 50% of the wingspan where height refers to main gear tire above the ground. In addition, one level-flight trim condition should be conducted out of ground effect (e.g., at 150% of wingspan).
(2) Shallow approach landing. The shallow approach landing should be performed at a glide slope of approximately one degree with negligible pilot activity until flare.
c. The lateral-directional characteristics are also altered by ground effect. For example, because of changes in lift, roll damping is affected. The change in roll damping will affect other dynamic modes usually evaluated for FFS validation. In fact, Dutch roll dynamics, spiral stability, and roll-rate for a given lateral control input are altered by ground effect. Steady heading sideslips will also be affected. These effects should be accounted for in the FFS modeling. Several tests such as crosswind landing, one engine inoperative landing, and engine failure on take-off serve to validate lateral-directional ground effect since portions of these tests are accomplished as the aircraft is descending through heights above the runway at which ground effect is an important factor.

Additionally, should you want to check a bit more, here is a link …
eCFR ? Code of Federal Regulations

[mnttech]

In re-watching the video (until I wanted to lose my dinner ) some points crossed my mind:
1. For each landing/departure, would it not be fun to know what the ground level TDZ, mid field, and far end (departure end?) reported winds were?
2. The same points 100 and 200 feet in the air.
3. The demonstrated cross wind component of each aircraft.
4. How many times the wind exceeded the aircraft component?

I’m not sure, but I do not think that FAR 60 discusses the runway random vertical deflection as shown in the video. Slope yes, deflection, I do not think so. I’m also pretty sure there is not a lot of data or design on what the wind does coming over the top of the buildings. All of which are going to be needed to improve the software.

The NTSB report on Continental Airlines Flight 1404’s Runway Side Excursion in Denver, Colorado on December 20, 2008 has the following recommendation:

Gather data on surface winds at a sample of major U.S. airports (including Denver International Airport) when high wind conditions and significant gusts are present and use these data to develop realistic, gusty crosswind profiles for use in pilot simulator training programs. (A-10-110)

I’m not aware that has happen as of yet.

I do think Part 60 should have discussed both the instruction of the instructor and the maintenance staff training, or at least made them required points within the SQMS.

Quick question for the current pilots: On a United States ILS runway with center line lighting, are the center line markings over the lights, or are they off to one side? And if off to one side, which side? Let us limit it to KSFO, KFJK, KDEN, KPHX, or some other major. I’ll answer the why in a day or two.

[AirRabbit]

Hi mnttech-

You might be interested to know that the FAA is about to publish a brand new notice that they are proposing some rather lengthy modifications to the Part 60 rule governing simulator evaluation/qualification.

It should be available in the US Public Register today or tomorrow, and if all else fails, certainly at the beginning of next week, and public comments are certainly invited.

This proposal would incorporate changes into part 60 that would either directly or indirectly address the following NTSB Safety Recommendations through improved FSTD evaluation standards to support the outlined training tasks:
- Stall training and/or stick pusher training
- Upset Recognition and recovery training
- Engine and airframe icing training
- Takeoff and landing training in gusting crosswind conditions
- Bounced landing training.

In general, the proposed changes to the technical standards would apply only to those FSTDs that are initially qualified or upgraded in qualification level after the final rule becomes effective. For previously qualified FSTDs used to conduct extended envelope, airborne icing, gusting crosswind, and bounced landing training, the FAA is also seeking comment on a proposed FSTD Directive that would require FSTD Sponsors to retroactively evaluate those FSTDs against certain objective and subjective testing requirements as defined in the QPS appendices and modify them if necessary to meet the proposed requirements.

[Goldenrivett]

Hi mnttech,

On a United States ILS runway with center line lighting, are the center line markings over the lights, or are they off to one side? And if off to one side, which side?

The painted line is to one side of the RWCL lights.
Google Earth shows KFSO 10R has the line to the RHS of the CL lights.

EGKK is similar, image of KDEN doesn't resolve the lights well enough.

[mnttech]

AirRabbit,

I see being friends of Dr. C has some advantages. I thought they could release a FSTD Directive without going through the US Public Register time? i.e. #1 on the visuals.

Goldenrivett,
Thanks, I was hoping a pilot would answer from experience… I’ll post over the week end my center line fun.

[AirRabbit]

mnttech -

As you probably know, the definitions section of Part 60 contains the following:
FSTD Directive — a document issued by the FAA to an FSTD sponsor requiring a modification to the FSTD due to a safety-of-flight issue and amending the qualification basis for the FSTD.

Perhaps one of the key words used in this definition is the word “requiring” – which certainly carries a note of imposing some thing on some one (individually or corporately) and, apparently, doing so “by right and authority,” as that is generally and, I believe, legally, the understood application of that term. Additionally, it has been, and continues to be, my personal impression that the FAA (or any other regulatory authority for that matter) would not be allowed (either by statute or internal standards) to impose new regulatory requirements, or change existing requirements, without providing the public an opportunity to comment on, even question, the proposal.

Of course I wouldn't have any first-hand knowledge, but I could easily believe that claiming to be a “friend of Dr. C” might sound like the comment one might offer through the peep hole provided in an overweight and demonstrably thick wooden door, located at the end of a darkened backstreet alley, when one desires admission to a mysterious establishment in order to engage in some likely illicit and/or potentially compromising business or activity. I, for one, wouldn’t necessarily go around uttering that statement, even whispering it.