Flight Dynamics - The Cold, Hard Facts (and some Opinio...

[Guest Ron Freimuth]

>>The Lone Gunmen realize this statement is "hogwash", so>does>>everybody else, >>Marc,>>I wouldn't be so sure about it. I work at NASA Ames and we>develop complicated aircraft prediction-trajectories for the>product called CTAS (now tested at DFW Center) - an ATC>software tool. Hmm, I worked at Ames over 30 years ago. But, not directly in aerodynamics. McFarland was worked at Ames. He wrote some general flight model code around 1970 that I bet was worked into many simulators since. Maybe even MSFS. I see R.D. Jones also worked at Ames, but not when I was there. Back then, I wouldn't have known him from Adam. ;) Of course, I was familiar with the '40X80', and most of the other WT's. Also, with the Manned Flight motion simulators. To generate 'scenery', one simulator had a runway painted on a big wood board. Little 3D buildings, etc were glued on the board (it was in a vertical plane). A TV camera was servoed to the simulator and the pilot viewed a TV projection of the physical model of the airport.> .... You would think that>NASA-FAA would have no problem getting such data from Boeing>or Airbus. The reverse is true. It is extremely difficult to>get them and what we have was derived indirectly. I heard on>multiple occasions that aircraft manufacturers for reasons>known only to them are very reluctant to release the data. True, McCormick says the same thing in his 1995 engineering text. Dr. McC had to 'fit' SFC to a polynomial series to give a graph of SFC vs thrust for a JT9 he analysed. IOW, he reverse engineered the graph, somewhat as I do for my MS flight models. Dr. Lowry, author of the AIAA "Performance of Light Aircraft" owns a C172, and measured the wing and prop before applying theory to get detailed results. Dr. McC owns (or owned) a PA-28 and his text has a lot of engineering data on it. Incidently, there is an accepted set of Stability Derivatives for the C172/182 airframe. They appear to be exactly what McC's analysis programs generate. Too bad MS didn't use those SD's in its more recent Cessna models, I did and they fly much better. Actually, I think the FS98 C182's inherited the good SD's from BAC, but MS managed to mess them up more and more as new FS versions came out. I figure if detailed drag curves were available, an engineer familier with airfoil design might be able to tell that a certain approach was used by a competitor. With either good or not so good results determined by years of post design flight experience. Similar for turbine details. Such information on the competition is very useful, since it can demonstrate that subtle variations attributable to some airfoil design approach (not clear even if one measures the wing and other profiles) has or has not been used by the competion. Big airframe companies are willing to spend a large amount of $ to reduce the Drag Count by 1.0. Over millions of miles of airline flights, that will save a lot of money on fuel costs. One "Drag Count" is a change in CD of 0.0001. Out of a total of perhaps '220 Counts'. Incidently, AFSD displays MSFS drag compenents of Cdp and Cdi with about that resolution. The trick is to get those that change to have the right form. >For some other work we also need airline specific data - for>example which flight # correspond to what aircraft N#. Again,>airlines flatly say 'no' to such requests citing all kinds of>reasons.>Michael J. However, FM's and other airline info always manages to get scanned and forwarded to people like me. ;) Much is also posted on WEB sites. I have over a dozen hard copy scans of 727 tables, and more as image files. The 727 is obsolete, and I doubt the airlines would complain anyway. I've also had commercial pilots noting flight pitch, panel readings, etc. for me during their flights. --- Other 727 data was recorded off the D-A converters of a FM simulator. As I gather world wide contacts I accumulate some information that is generally considered propriatory. I have a Boeing Performance Engineering manual on one of their AC maked 'propriatory'. They only sell it to owners of the jet transport. It gave me an idea of typical LG drag, Mach effects, etc. However, I more or less reverse engineer total CD from FM tables. Since I get good correspondence between PPH and the tables over a wide range of FL's, Mach Numbers, and weights I think I create reasonable models of AC drag. In particular, "Lift dependent drag as a function of Mach number". SFC is constant in MSFS so variation in turbine efficiency doesn't come in. But, that is nearly constant in cruise conditions anyway. Typically, Cruise PPH has a Std Error around 2%, with the extremes about 5% off. Over a range of cruise FL's, weights, and Mach numbers (assuming I have enough FM data). Climb and descent is problematical, since SFC decreases at lower speeds in real turbines. At any specific flight condition one can read Specific Range as accurately as desired. It is simply TAS/PPH. AFSD displays it. I've also programmed 'Specific Range' into my Jet and Prop test gauges. I was recently seeing numbers like 1004.45 nm/klb when looking for deviations between FS2K2 and FS2K4 drag. I found the difference between sims was 0.05% in my C208 (tested at one low and one medium speed) when it had been **correctly updated** for FS2K4. I'd consider that neglible. Blame the difference on 80 lbs weight variation. Averaged on a long flight at multiple FL's, I expect block fuel burn to be within 2% of the real AC. However, this requires so much testing I have not verified this in general for any of my jet transport models. I saw the 2% error for my 747 when flying it against an offical FP from Chicago to Tokyo. Starting at FL290, last leg was at FL 410. I could use FSNav to set the path and altitudes back at that time. Speed must be very close, I set Mach within 0.003 or closer but actually work with TAS, since that is also in the FP tables. I have also used the BADA data which gives climb and cruise estimates for a wide range of AC. However, I know their drag model is limited and doesn't consider increase in drag past Mcr. I expect to have engineering data on Rutan's Voyager 'from the horse's mouth' one of these days. Also, data from the engineering department of a SEL company. One runs into a wide range of aviation professionals right here at AVSIM and over time I am obtain more and more arcane information. -RAF

[Guest]

Thanks. It looks great. I can't afford $40 (Cad) right now though :)Mike

[Guest]

WS,Thx. More in my price range at the moment .Mike

[michal 44]

Ron,Predicting aircraft trajectories calls for a slightly lower simulation fidelity than doing full 6-degree flight dynamics. But we still need basic power/drag/lift equations as well as good engine data. CTAS stands for Center Tracon Automation System. You can read more about it here: http://ctas.arc.nasa.gov/I used to work on CTAS a couple of years ago. Now I am onto something else - new method of spacing flights in congested airspace.Michael J.http://www.reality-xp.com/community/nr/rsc/rxp-higher.jpg

[badderjet 181]

Nice post...

[robert young 2,897]

And here is a cast iron classic reason for the concept of copyright.RY