Originally Posted by misd-agin
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.
While I completely understand and agree with most of your comment, I think the term “horrible” really should be taken as a range of “acceptability vs realism.” Of course, no simulator – at least at this point – can do all the things an airplane can do … and probably the most immediately recognizable is the “you-are-about-to-die” aspect. Although, I can say that I’ve been on-board a simulator when the building in which it was housed was directly hit by a bolt of lightning that actually shoved the simulator’s 6-axis motion system capabilities to the maximum extension in each direction and probably exceeding at least some! Scared? You bet! The sim cab stopped in about a 30 deg nose low and as close to a 90 deg left bank that could be achieved with the 54 inch jacks. First and only time I’ve had to use the escape ropes to get out of a simulator!
Originally Posted by Bergerie1
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.
Good question … I wish I had a solidly good answer … so until one comes along … I’ll use this one. But, I warn all readers of my propensity to use “a lot of words,” so bear with me, please.
Originally, back when training was conducted almost exclusively in airplanes (yes, there were things called “simulators” back then but they were extremely limited and provided good training but only relevant to cockpit procedures and systems operations) the idea of conducting flight training was a serious issue … particularly with larger, air transport airplanes, and most particularly with stall training.
We all know that an aerodynamic stall and recovery does not necessarily result in a spin – but most instructors had that potential stuck firmly in their memory banks – and some students were prone to grab the aileron and elevators controls and literally “kick” and hold the rudder, doing both in such a manner that would rival any gorilla’s determined grip on any tree branch! It didn’t happen often, but it didn’t take many times to prepare an instructor to do almost anything to regain control of the airplane – Right Now! Of course a spin is a spin, and recovery from spins may differ from airplane to airplane … and as you would easily understand, the larger the airplane the more … um … interesting … the experience. Instructors were prone to be very observant and very demonstrative in describing and demonstrating how to get into an approach to stall and how to recover from that condition. No one wanted to get the airplane into a stall.
Recovery from stalls in most airplanes was, as you would immediately recognize, almost always accomplished by applying the proper rudder to center the ball (zero the yaw), applying elevator controls to decrease the angle of attack (sometimes this meant a rather large control movement forward), and aileron input to put the airplane in a zero bank – all simultaneously. With a slightly (sometimes a bit more than “slightly”) nose-low attitude, the power was advanced symmetrically. The result was generally a proper and effective recovery. Well … that is essentially, if not exactly, the procedures that were to be applied to recover from any approach to stall … where an “approach to stall” was defined as the initiation of any installed warning (warning horn, warning light, or stick-shaker), or the initiation of the elevator/airframe buffet. This practice became the routine for training in an airplane– take the airplane just to the approach to stall and teach these same, correct, recovery procedures – emphasizing that anything that went beyond the simple approach to stall warning or any fully developed stall, should be responded to in exactly the same way. This was the routine for airplane training – and it did not include training to recover from fully developed stalls. When simulators were on the rise, and the clear direction the training industry was moving … toward the greater use of flight simulator … the fact that full stall training in the airplane did not include anything beyond the approach to stall, the acceptance of the use of simulators to teach recoveries from the same approaches to stall was not a huge leap – and was, for quite some time, considered to be quite adequate.
Another problem that exacerbated the situation was the fact was that engine acceleration was directly coupled to the throttle position, and there was very little simulation of airflow taken into consideration for the computed level of thrust generated by the engines. If the pilot pushed the throttles forward, the engines were programmed to produce the appropriate amount of thrust based on that throttle position, regardless of the attitude/position of the engines. Unfortunately, in some simulators, slowing to an approach to stall required an increasing pitch attitude to maintain the established deceleration rate. Even those who would accept a slight rate of descent to be generated also increased the AOA to continue to airspeed deceleration. The relative airflow that would have gone into the engine with that specific AOA was not calculated or at least was not programmed into the simulator. This situation, in combination with the edict that a “minimum loss of altitude” was expected during the recovery, resulted in “perfection oriented pilots” (and I’m sure you know the kind of pilots I mean) to want to keep that pitch attitude throughout the recovery. You and I know that an airplane with a relatively high pitch attitude or at least a relatively high AOA, will not accelerate as quickly when the power is added if the pilot maintains that attitude or that AOA as it would if the attitude or AOA is reduced. Unfortunately, because of the error in the simulator programming, the pilots using these simulators found that by holding the pitch attitude (perhaps even increasing it slightly to aid in the airspeed reduction) and simply advancing the throttles, the airspeed would increase, and regain flying speed in a relatively short time.
Unfortunately, THAT is wholly unrealistic. It was an error in the simulator programming, BUT it was what was programmed into at least some simulators. Fast forward to the ABX DC-8 crash in Narrows Virginia on December 22, 1996. The pilots on that airplane (up for a post maintenance check) attempted to do exactly what they learned in the simulator – and it cost them their lives. They refused to lower the nose, and kept attempting to “fly out of the top of the stall.” With the attitude they had, the airflow into the engine intakes and the throttle position calling for “more” simply resulted in engine compressor stalls.
Despite the publicity of this 1996 accident and the conversations regarding stall training, there wasn’t much that changed in simulators and not much that changed in the way simulators were used to teach recoveries from approaches to stall. Again, fast forward to 2009 – the Colgan Air, Dash 8-Q400, crash on February 12, 2009 in Clarence Center, New York. As a result of this accident more scrutiny was aimed at the way this crew was trained, and the realism that simulation can or cannot provide in that portion of the “flight envelope.”
There had been concerted efforts proposed and some actually begun over the last 8-12 years – but without knowing exactly what was at risk and/or what could be achieved, a whole-hearted effort (as costly as such efforts have been and are likely to be in the future) pressing forward and accelerating the efforts were difficult to achieve. So, reasonable facsimiles of such a robust effort were initiated by several organizations … two of which were the US FAA and the UK’s Royal Aeronautical Society (RAeS). Since the Colgan accident and the publicity that tragedy generated – particularly the publicity focused on the US Congress – the FAA has taken several serious steps forward – some to initiate and some to advance or increase some programs already begun, to examine simulation and it’s potentials. The same kind of discussions have been taking place at the RAeS – and there has been considerable focus directed to these and associated flight training and simulation issues. As a result, today there are several on-going, multi-corporation efforts underway that are looking specifically at how simulators can and/or should be used to train for airplane situations that are at or outside of the normal flight operational envelope. Certainly, an aerodynamic stall would be considered outside of that normal envelope – but approaches to stall – where the recovery is initiated at the first indication of an approaching stall, is very likely at or approaching the boundary of that envelope – and, significantly, not beyond that boundary. The question then becomes how far past the normal envelope boundary can we go, should we go, and with what accuracy, to allow the programming of simulators? How does one gather data in an area where there is no consistency; where there are no accurate predictions; where airflow patterns are “random;” and all the rest of the issues that most pilots know in general concept … but few, very few, know what or how to go about defining, or measuring those factors, or know how to structure that data into a reasonably accurate aerodynamic model that reasonably represents an airplane and put that model into a computer-run, firmly bolted to the floor, reasonably accurate representation of a familiar airplane cockpit!
Well, thanks to the diligent efforts of some of these groups of professionals – today, I know we have at least one, and now, I understand, a second, transport category airplane simulators (the first one IS and I believe the second one is ALSO a B-737) that have an aerodynamic model installed that is accurate enough that the several test pilots (2 or 3 on the first, and likely up to 7 or 8 on the second) who have flown those simulators, have reported that the simulator performs and, critically, handles, as much like the airplane (the B-737) throughout the aerodynamic stall entry, the actual stall, and the stall recovery, as any anything they have seen. As an example of the competency of these pilots, one that I witnessed personally, when an interested observer quizzed one of these test pilots how far he had personally taken that airplane into the actual aerodynamic stall … he answered, “a 3-turn spin.”
At the moment, I am not aware of precisely where this program is and how much additional work on data gathering, data reduction, verification, and validation has yet to take place; nor am I aware of precisely how close we may be to the commercial development of such models. As I’ve indicated, here and elsewhere, I am now retired and working as a generalized flight training and simulation consultant – and therefore do not have the same connections to these programs as I have had for the last several decades – although I do retain some rather close friendships with many who are directly, intimately, up-to-their-necks, involved. We also have to determine what kind of rule or regulation would be necessary or appropriate for the incorporation of this aerodynamic model (these models – presuming there are 2) into airline simulator use – including the requirements that would, should, or could be leveled on the pilots using these simulators – if any. Presumably, if the process is as good as it appears to be … future questions might be, will there be any follow-on development of aero-models for other airplane types/models/series? Also yet to be determined is the cost of this whole package and whether or not a regulatory authority can or should mandate that airlines purchase and incorporate these aero-models into their fleet of simulators – or if there should be some kind of governmental reimbursement scheme. Also, I am not sure if there would be a restriction on the level of simulation that would be required to accept these kinds of simulation programming scenarios – however, it makes sense that the level of simulator would have to be relatively advanced.
As I said earlier … I believe we haven’t seen anything like we’re likely to see in the future with respect to the use of competent simulation – and personally, I’ve committed my efforts to do all I can to encourage all who might be interested to join the effort – to any degree – to any level that makes sense. Personally, I’d like to see colleges and universities who have active aviation programs, expand those programs to address the kinds of things that research will help resolve with respect to airplanes being satisfactorily simulated with a computer program and a good looking model/replica of the airplane cockpit. Simulation is very likely the very best thing that has or could have happened to the aviation industry – and it will need a lot of professional guidance, observance, critical evaluations, suggestions, and who knows what else to take the maximum advantage of what I think is a very bright future in this area.
I hope I’ve answered your questions … but, more so, I hope I generated a newly charged interest in this particular aspect of aviation. Thanks.