Finite Element Analysis help needed

[rv8ch]

In order to get my airplane registered here in Switzerland, I need to prove to the local FAA-type people that it won't fall out of the sky. Unfortunately, hundreds flying elsewhere don't seem to make much difference to them. Fortunately, Van's has supplied them with almost all the information they need to be happy, with the exception of my engine mount, and the gear legs.

The only things they accept are a "FAR 23 drop test" or a "finite element analysis". I don't really want to do a 9g drop test on my airplane, and while the suppliers of the engine mount and gear legs (Eggenfellner and Grove) claim that their stuff will pass this test, they don't have any documentation to back up this claim.

So, I'm left with the finite element analysis. I have a rough idea of what FEM is, but I was wondering if there is anyone on the list that could help me with the following questions:

1) do I have any chance of downloading software, figuring it out, and performing the calculations myself? I have an engineering degree, but we never studied this. I doubt I could still solve a partial differential equation to save my life, but I do remember how to add and subtract, and I *can* work a mouse.

2) if it is possible do do it myself, do you have any suggestions for good software to do it?

3) are there companies or freelancers that could do this for me, and if so, about what would I expect to pay for this service? Any recommendations?

Many thanks for any hints or tips.

[Radomir]

4) N-register it, fly Mel over and come out cheaper?

<I'd say something else about those regs, but I'll refrain to keep it PG rated>

[AlexPeterson]

Mickey,

Bummer about the rediculous requirements there... can't fight city hall though.

FEA is generally not something that can be done casually, even by engineers, as operating the software to do it is not simple. Most FEA software is quite expensive ($10k and up), as there are usually very few users. If there are any free programs, I suspect they are worth what is paid.

I do some of it, and can certainly have a look at the gear legs, as they should be quite straightforward once you send me the geometry. The motor mount may be another matter altogether, as the geometry and loading conditions could be challenging. Have they been at all specific with what they expect with an FEA analysis? Do you impart some load to the main gear, and look for peak stresses?

If you have to pay someone, take a deep breath. Those guys do not go cheaply.

[jcoloccia]

My guess is that it would be cheaper to build another RV with dead weight for your engine and stuff, and drop test THAT than do an FEA on your airframe. I can't imagine that their requirements are so stringent. There's gotta be an easier way.

[Kahuna]

My good friend, fellow formation junkie, retired F-15, now lyco powered rv-4, Bob "Subob" Goodman of Atlanta Georgia does this for a living and has done this on his custom Subaru engine mount (back in the ole days when he was running a Subaru in his RV-4, hense Subaru Bob is not "Subob"). I dont think it would be that expensive if the designer would provide the drawings.

You can try him here Goodex@aol.com and I have sent him the thread.
Best

[jtrusso]

I agree with the above posts that it's not really feasible to do this yourself if you don't already have the software and are very familiar with it. I'm a mech engineer and the two software packages I'd use to do this type of analysis, NASTRAN or ProEngineer Mechanica both run upwards of $20K a license. I've been doing this for 10 years and have no idea how to run NASTRAN, and a fairly good understanding of ProE, but I've got over 12,000 hours of ProE work under my belt.

I'd question the authorities by exaclty what they mean when they say they want an FEA. If you have a background in mechanical engineering I would think you could perform a basic analysis of both the gear legs and engine mount using excel. Does it really have to be a finite element analysis? You could do a static analysis and then an dynamic impact analysis, simulating the engine as a point mass on the frame. After all, a drop test is really a impact test afterall. There are several good text books out there that cost an order of magnitude less than the software that could explain how you'd set up such an analysis (http://tinyurl.com/mtob4, http://tinyurl.com/h6vme, http://tinyurl.com/pu44r) if "the man" will accept it as a substitute for the FEA.

[mlw450802]

Originally Posted by rv8ch

In order to get my airplane registered here in Switzerland, I need to prove to the local FAA-type people that it won't fall out of the sky. Unfortunately, hundreds flying elsewhere don't seem to make much difference to them. Fortunately, Van's has supplied them with almost all the information they need to be happy, with the exception of my engine mount, and the gear legs.

The only things they accept are a "FAR 23 drop test" or a "finite element analysis". I don't really want to do a 9g drop test on my airplane, and while the suppliers of the engine mount and gear legs (Eggenfellner and Grove) claim that their stuff will pass this test, they don't have any documentation to back up this claim.

So, I'm left with the finite element analysis. I have a rough idea of what FEM is, but I was wondering if there is anyone on the list that could help me with the following questions:

1) do I have any chance of downloading software, figuring it out, and performing the calculations myself? I have an engineering degree, but we never studied this. I doubt I could still solve a partial differential equation to save my life, but I do remember how to add and subtract, and I *can* work a mouse.

2) if it is possible do do it myself, do you have any suggestions for good software to do it?

3) are there companies or freelancers that could do this for me, and if so, about what would I expect to pay for this service? Any recommendations?

Many thanks for any hints or tips.

Mickey,
I don't know what your time frame is here but the short answer to your question is that it is unlikely that any free software will give you realistic answers and the learning curve is pretty steep to get reasonable results.

The major effort is in getting a nice representative solid model of your system then meshing it appropriately for the results you want.

There are several packages that will solve drop/impact type problems and the real world answers are different from the static 9g test that you can probably do with your calculator and a statics textbook. The differences are the inertial loads that are hard to handle because the truss geometry of the engine frame changes as it deflects and the dynamic 'impulse' is hard to estimate due to the cushioning effect of the inflated tires etc. That is where the FEA comes in; all those individual components are calculated in concert with one another.
I use LS-Dyna and ALGOR MES in my work with the nod to LS-Dyna in terms of capability.

Solving your problem would be a piece of cake once the correct model was built (true of any FEA problem)and it would only take some cpu time to get there.

I would be willing to work the problem if I could get someone else to provide the model but unfortunately, I just don't have time to do the model creation. I have followed your site for quite some time and would be willing to give back some of what I have received from all your input and experiences.

Generally, a rough estimate of modeling the system might be 40+- hours (if all dimensions are readily available) plus another 20 debugging.

-mike

[ferret]

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. Any FEA results have to be verified by loading a test article and measuring the strains with strain guages to verify the model predictions. Takes alot of tweeking. I don't think you should attempt the FEA approach.

[Ironflight]

Originally Posted by rv8ch

....I have an engineering degree, but we never studied this. I doubt I could still solve a partial differential equation to save my life, but I do remember how to add and subtract, and I *can* work a mouse.

.

I feel your pain Mickey...yet I'm afraid I can't give you anything but encouragement. I too am allegidly an engineer, but school was a long, long, time ago! I came across a folder of some space pointing software I wrote when I was a young flight controller - not only don't I understand the code I wrote, but the spherical geometry stuff I used also looks pretty mysterious...and I'm not even sure I coudl do the arithmetic anymore!

On the plus side...the young Engineering Coop Students I had working with me on my flight test data were whizzes at running the latest Mathlib, curve-fitting software and stuff, and could really understand C++. But when they had to run Carl Morgan's early versions of the GRT Decode program, they were stumped - they didn't know how to run programs from a command line in DOS! So maybe there is hope for us old guys after all...

Now there's an idea! Have you got any connections with your old engineering department? Coops or Interns where you work? They might very well have the skills you need, you have a project to complete, they'd get real-world experience....mutual benefit?!

Or just hoist it to the rafters and drop the thing...

Paul

[David_Nelson]

Hi Mickey,

I did a search on Freshmeat.net and came up w/ a couple of s/w packages.

http://freshmeat.net/search/?section...ent%20analysis

HTH,