Doable but you might need an engineer
ferret said:
FEA is more art than science. Getting the boundary conditions are very difficult. You need a pre and post processor like ProE or PATRAN, and then the analysis code like NASTRAN. Takes a lot of tweeking. I don't think you should attempt the FEA approach.
IT is not that bad and you can use simple programs, but is that acceptable? The ironic thing is kit planes don't need to meet FAR's. There is no guarantee it will meet FAR23?
I do agree with you all about FEM. Stress analysis, including FEM was my job for many years. The idea of getting an engineering student to do it, not bad. Still you need to get the data together.
First how bad is the drop test? I think 9 g's is may be over stated. I hope. A quick look at the FAR's says you need at least 9.2 inches but no more than 18.7 inches. Also weight is offset by wing lift, so you are not dropping at full fuel+payload wt. If it was 9g's I would never do that. That with out calculation I think would "LEAVE A MARK" or bend the plane or gear. At 9's at just empty weight is an equivalent of over 1500 lbs on the nose and over 4000 lbs each main gear! Ouch! At gross fwd / aft CG's respectfully that's about 4000 lbs nose - 7000lbs mains and one heck of a large deflection.
Interesting enough the actual G force from the impact is hard to calculate because you don't know how long it takes the airplane mass to decelerate to zero. With some guessing the G force is about 3g's. You have to include the tire deflection and of course the gear deflection (bending). The tires will scrub and absorb the impact. Its a really interesting engineering problem to analyze DYNAMICALLY with the rebound, thus the reason for actual test. That dynamic solution analytically is of course more complex than a static solution. Better find out what they want or need before you do anything. I have doubts about what they are asking?
DROP TEST?
http://www.risingup.com/fars/info/part23-725-FAR.shtml
http://www.risingup.com/fars/info/part23-723-FAR.shtml
other gear FARS:
http://www.risingup.com/fars/info/part23-726-FAR.shtml
http://www.risingup.com/fars/info/part23-727-FAR.shtml
http://www.risingup.com/fars/info/part23-473-FAR.shtml
23.479, 23.481, 23.483, 23.485, 23.493, 23.497, 23.499
http://www.risingup.com/fars/info/part23-499-FAR.shtml
This gives the hight and effective weight. It does not sound too bad. At 18" you would take 0.3 sec or speed 9.5 ft/sec. However I am not crazy about dropping the plane. Also how do you set that rig up to drop it? It would be an effort. I lean to analytical solution.
Hey its experimental why the FAR's? Again there is no guarantee.
I never made a 9g landing in my RV, and I have a g meter (that is my story and I am sticking to it). Also to be correct one drop does not design a gear system, since there are drag, spring back and side loads. Its a start but I have no idea what they want? Is it written down somewhere. To have a full analytical and or scale gear test is way more than I think you want or need to undertake (I hope).
ANALYSIS? FEM
First you need the software to do the FEM. Full bore FEM programmes ones that use NASTRAN or PATRAN code, are many thousands of dollars for just the basic elements. Basic ones can be a few $100. Excellent programs with a basic CAD interface can be inexpensive. Here are a few free one I googled. Have no idea but could do the trick?
http://www.lisa-fet.com/index1024.htm (it looks like a full 2D/3D Fem with a basic pre/post processor, even w/ plates, shells and 8-node solid elements. Just downloaded it, never used it. It's a demo limited to 1300 nodes.)
Very simple 2D and 3D truss program:
http://www.cuylaerts.net/
Some free/demo FEM programs:
http://www.freebyte.com/cad/fea.htm
Not all use NASTRAN code (the Cadillac and widely accepted code by aerospace world wide).
You really need to know "mechanics of materials" to use one effectively; as ferret said you need to choose the boundary conditions, element type, properties and how to apply the loads. I have done a lot of FEM in my engineer days, not hard but takes judgement and time. You need to have all the geometry and material properties. There are handbooks with the material properties. The fancy programs fill in the blanks for you with a material data base. "THEY" may not accept non NASTRAN code?
In your case the FEM would be like a 8-year olds "Stick-MAN" drawing compared to Lenardo Da vinci's "Mona lisa". You would just need to model the gear and local support structure. The rest of the plane would be a few dummy elements (truss/beam
) to connect the dots.
How to model?
Elements, Boundary conditions, if not chosen carefully they can throw the results off. The gear has all 6 degrees of freedom fixed at the support. The other end is free or cantilever. I don't know but assume they want some knowledge the support structure will pass mustard. The airframe I assume they are happy with? Not sure what you told them to make them believe than but not the landing gear?
In the case of a RV-7a that is some simple truss (engine mount or main gear leg fittings bolted to the firewall or wing spare structure). The rest of the airplane is modeled with infinitely stiff dummy beam elements, which ties it all together.
To model the gear legs themselves you can do it
two ways.
One is solid FEM elements of the gear legs, which is simple if you have the definition or geometry and material properties. Most FEM programs can take a CAD drawing of a part and "mesh" it. The
second way is represent it with a standard FEM tapered beam elements, which saves you from modeling the gear legs with solid elements.
I would be glad to help, but I can't volunteer or promise to do number crunching. I always thought someone should do a solid FEM model of the RV "A" model's nose gear and see what was going there. I just can't believe someone does not have a FEM model already, even a full airplane model.
FEM software is getting cheaper, more prolific and easier to use with graphical interface than they where 10-20 years ago.
Of course PC computing power is what drove this. FEM analysis once only ran on mainframes. Even today solutions of a larger model can take hours or all night to solve on top of the line PC's working in parallel. The tricky thing is as the gear bends the geometry changes, a lot! That is the issue with the nose gear, it deflects. There are special elements to solve for this "non-linear response" or "large deflections", but you can make some assumptions to simplify this. We assume we remain in the elastic range?
Stress Analysis:
The last way is classic stress analysis, i.e., pencil and paper. This is what they did before FEM and wide spread computer use and is still very accurate, especially with simple structure like this. FEM use to be used only for very large analysis jobs. Now it is common to model simple structure.
Some times its easier to do a quick and dirty model (FEM) even for simple structure. Just depends on how adapt the engineer is at hand calculations using standard stress analysis techniques. Good hand calculations are as good or better than FEM in my opinion many times. WOULD THEY (your FAA) ACCEPT IT? I know the FAA in the US would be fine with it. The thing about hand calculations with "classic" formulas is its easy to check.
If you gave me
all the data I could try to do a quick crunch if the number by hand. No promise. It might be acceptable? It would show the max stress (guess). I would compare that to the limit allowable (assuming you gave me the steel alloy and heat treat). If the stress is less than the elastic limit, than you are good to go. I see NO reason why they would not accept that.
FEM will simply returns the stain which is stress when multiplied by the modulus of elasticity for the material (E). This is called Hooke's Law.
Even though the geometry changes under load, i.e., the gear spreads out, it is statically determinate. If they would accept a page of stress analysis I might be able to dust my mechanical pencil and HP41C off. I would have to calculate the stress for side load, drag and vertical per part 23 and just crank the numbers.
DECIDING
If you go analysis regardless who does it you need to give them all the info.
You would assume if the fuselage and wing are good for flight (by their estimation) so this does not need to be shown.
What did you show them to make them happy the wings stay on?
Therefore analysis needs to only show the gear(?), engine mount and the main gear fittings. Interesting a hard landing can break wings of some planes well before the gear, especially ones with tip tanks. A guy landed hard in a twin Comanche with tip tanks full and bent the wing. The gear was fine.
Write me off line if you want help. Can't promise to crunch the numbers, but I'll see what I can do. If we can get a FEM analysis document on file with your "controlling agency", with a range of weights, all RV'er could be "gear" approved in your country if they fit the approved weight and CG range. Why does this have to be done over for every plane?
Van is an engineer. He might have detailed stress analysis? At least they should provide you info on the gear material and we have the geometry. Best of luck. Again there is no guarantee our RV's meet any part 23 FAR. I think the case that there are almost 4000 flying, at least 1/2 with your exact gear configuration for at least 10-15 years should say something.
Cheers George