Modelling a carbon fibre chassis in FEA package

[mossy]

I am looking to undertake static analysis using an FEA package to determine the strength of our chassis under loading.

Has anyone successfully modelled carbon fibre with a core material within an FEA package?
I have no material properties or understanding of how to model the ˜sandwich' of pre preg carbon fibre either side of a core material.
Does anyone know how this is achieved or knows of any reading material I may be interested in or tutorials? anything, please

Cheers

Mossy

[mossy]

I am looking to undertake static analysis using an FEA package to determine the strength of our chassis under loading.

Has anyone successfully modelled carbon fibre with a core material within an FEA package?
I have no material properties or understanding of how to model the ˜sandwich' of pre preg carbon fibre either side of a core material.
Does anyone know how this is achieved or knows of any reading material I may be interested in or tutorials? anything, please

Cheers

Mossy

[SNasello]

You could maybe try and do some mechanical testing on a sample of the carbon-core-carbon sandwich. A simple 3-point bend test can tell you a lot about the mechanical properties.

[Tony K]

Well, testing a pre-made sandwich panel will give you the modulus of the structure, but it won't relate to strength as the geometry between the simple plate and chassis are a bit different. There are ways to do FEA on a composite structure, which can give you an estimate of your chassis strength and deflections, but there are a lot more bad assumptions that have to be made.

At the moment we're using a combination of Pro/Mechanica for the overall structure analysis and a program called FiberSIM for fiber orientation. CATIA has a few composites analysis add-on packages, and even without these specialized programs you can do analysis in just about every FEA package that I know of by defining shells on your part and then aligning the principal directions of the shell to that of the fiber orientation. If you're trying to do analysis on a pre-made part, and don't know the exact fiber orientations... good luck.

However the actual lay-up of your finished part is a huge factor in the determination of strength. As with all FEA simulations, the accuracy of your results degrades exponentially with the number of assumptions you make. With composite structures you end up with more assumptions than you'd like unless you have brilliant manufacturing capabilities with CNC fabric cutters, fiber placement machines, nice molds, autoclaves perhaps, and people with experience in composites laying up your parts.

I can tell you from our experience that translating from FEA to real world with composites is extremely difficult with a structure as complicated as a chassis, at least at the university level. We're in the midst of performing a series of torsion tests on our monocoque chassis, and even with our decent amount of composites experience and knowledge, the results are nowhere near what FEA came up with. There are companies out there that specialize in such things, ATR Group for example, and are very good at it, but it's because they have the correct tools, programs, and knowledgeable people with many years of experience in composites.

So I suppose to sum up my exceedingly long post, I'm not trying to discourage you from trying a static FEA analysis, just letting you know that it would be wise to perform some sort of physical testing on your chassis as well. Without testing, I wouldn't even feel good about estimating some sort of safety factor to translate FEA to the real world.

[Travis Garrison]

I'd agree with Tony, be very skeptical of your FEA results and focus on physical testing.

If you're going to do it right you probably want to start by understanding the laminate properties, which means reading:

http://www.amazon.com/Structural-Polymeric-Composite-Me...eering/dp/0824747178 (http://www.amazon.com/Structural-Polymeric-Composite-Mechanical-Engineering/dp/0824747178) (comes w/ some basic programs for laminate analysis)

or

http://www.amazon.com/Composite-Materials-Daniel-Gay/dp...2539?ie=UTF8&s=books (http://www.amazon.com/Composite-Materials-Daniel-Gay/dp/1587160846/sr=1-2/qid=1165978800/ref=pd_bbs_sr_2/102-8541812-8022539?ie=UTF8&s=books)
good examples in the back

I'm sure there are plenty of other good ones out there, but these are the ones I can vouch for http://fsae.com/groupee_common/emoticons/icon_smile.gif Also, look into Racetech's series on composite chassis design.

For analysis there are some older threads, so start with a search of the forum. Ansys and Nastran are the most common composites capable programs, MSC's SimDesigner is a more user friendly Nastran built into Catia so theres another option. No matter what you choose you'll still need to test your materials physically before going too far.

If you're just looking into chassis concepts I would start by assuming quasi-iso materials (conservative approach) and analyzing with an E somewhere around 7-10 MSI, use shell elements, and for weight calcs call it ~.05 lbs/cu inch. This shows about where you should be able to get with a bit of a shotgun approach (plus it will force you to make use of bulkheads rather than coring over problem areas). You should be able to do better, but often you won't on your first go round simply because it's hard to get a handle on all the different things that can go wrong between analysis and the real world.

Once you've got some laminate properties try asking around with people from teams that have done this a couple of times, most FSAE teams share pretty willingly, esp 1 on 1.