FEA (Orthotropic material)

[HyeZeus]

Any recommendations on FEA software for othrotropic material.

Or comments and recomendations.

Other than Algor, NeiNastran, MSC Marc, ASIC.

It is For a Honeycomb sandwich Graphite part.

[RKemmet]

ansys ls dyna pretty sure should do it.

[Matt Gignac]

No need for lsdyna, regular ansys should cut it. Shell 99 is probably the easiest to set up, although the run time will be longish. Shell 181 (i think, need to double check) has only 4 nodes per element, as opposed to 8 for shell99, so it will solve quicker, but maybe less accurate.

For the analyses I did, I used a linear isotropic material for the honeycomb (not actually true, but for my purposes this suffices) and a linear orthotropic material for the carbon

Ansys workbench won't do layered elements though, so you'll have to hack it with the shitty non-user-friendly version.

Matt Gignac
McGill Racing Team

[HyeZeus]

so you analysed each layer seperately right. How did you calculate your combined torsional rigidity?

[Matt Gignac]

This wasn't for FSAE, it was for my composites design class. Basically it was just to make a carbon fiber and honeycomb laminate, get stresses in each layer and failure criteria, etc.

In the Ansys postprocessor, you can get the stresses in the fiber and matrix direction for each different ply, and if you have the strength numbers put in, you can also get the safety factors for each ply for a few failure criteria. I think it's possible to get inter-ply shear stress too, but I haven't figured out how yet.

I gather from your post that you want to do the analysis of a chassis. For this, I think the best way would be to draw surfaces in your favorite CAD program, and lines to represent the tubes. To start off at least, it's probably best to not put any curved surfaces, just flat surfaces to get an idea of the laminates you want for each chassis panel. Export it as an IGES, which is quite easy to import into ansys. From there, you'll need to choose your element types (shell99 for laminates, some kind of beam element for tubes), make some real constants for each different laminate type and each different tube type, define your materials (you'll need 3, composite, honeycomb and steel), assign a real constant to each surface or line, mesh everything how you see fit, apply some loads and constraints, and give'r.

An important thing to look at is how your tubes are attached to your composites. If everything is bonded, just go merge all and you ought to be good. If you have some kind of bolted connections, you'll need to look at only attaching certain nodes, which i don't know how to do.

Anyways, if you want, i have a tutorial about composites in ansys, its pretty useful for getting the basics down.

Matt Gignac
McGill Racing Team

[TG]

Matt,

could you please send me the tutorial for ANSYS, I'm very interested in learning this stuff. I've just got a copy of it from the university and trying to figure it out. My e-mail is asuglax AT yahoo DOT com.

Which failiure criterion are built into ANSYS? Tsai-Hill? Tsai-Wu? or maybe just maximum stress or strain? Thanks a lot.

[HenningO]

It'd be great if you can send it to me as well:

enningo (AT) gmail.com

[John_Burford]

Matt

I'm curious if you have done any comparitive studies between an isotropic shell element instead of the linear orthotropic shell. In other words is the additional detail worth the added effort. If you are designing the Boeing 787 or Lockheed JSF of course it is. But for a mat lay-up on zeros and 45s (like most race car chassis) is an isotropic material good enough within some percentage. I'd love to hear from anyone with some practice experience. Thanks.

John Burford

[Matt Gignac]

Can't say I have. All the laminate optimization work i've done was for a composites course with only software calculations, so we see all kind of ridiculous plies.

For a practical lay-up though, I'm thinking a quasi-isotrpoic symmetric lay-up would be best, so you could maybe assume a certain percentage of fiber modulus, depending on the number of layers. HOwever, I think the main attraction ogf quasi-iso laminates is the reduction in inter-laminar sheer stresses, which would make classic laminate theory more accurate.

Fpr a typical fsae chassis though, I think it's pretty easy tp get the general fiber orientations you need for each panel. For instance, you can expet a lot of +/- 45 plies, as these will take the most shear forces, and maybe a whole bunch of 0 plies to take care of buckling and side impact loads. And if you awnt to minimize inter;aminar stresses, you'd probably design the rest of your laminate to be quasi-iso.

Long stpry short, I'm pretty drunk, but the main idea is to identify the direction of the major forces in each panel, put a whole wack of fibers in that direction, and then arange the rest of the fibers to minimize shear stress between layers.

If it were up to me, I'd do isotropic analysis to get a good idea of ways to arrange the various panels (more importantly to see if it's stiffer than an equivalent steel tube triangulation) , and to see the general forces in each panel, and then do some laminate analysis of each panel to optimize the stiffness, and make sure it meets the strength requirements needed (to meet side impact rules and such).

Matt Gignac
McGill Racing Team

[LD]

could you please send me the ANSYS tutorial too.

ailijic@eden.rutgers.edu

[HyeZeus]

me three. send me one too. hyezeus@aol.com. I need to get my hands on this software too. Ill figure something out. Neinastran allows for a 30 day full version trial.

[DJ Pohlit]

Matt, I designed the composite chassis for the Virginia Tech FSAE team last year, and I'd still like to play around with doing some more advanced FEA on the structure. Most of my design was based off of simple sandwich panel analysis and a lot of physical testing, but I'd like to take a look at some more advance FEA possibilities to help with future teams. Any chance you could send me the ANSYS tutorial as well? My email is dpohlit@vt.edu.

[Ppada]

I Really believe that this site has some tutorials that can help out, some of you. And of course if you work it more you should be ok for Ansys. BUT i think i have found some errors on these tutorials.


www.mece.ualberta.ca/tutorials/ansys/index.html (http://www.mece.ualberta.ca/tutorials/ansys/index.html)

[Enghamed]

Hi
please send me the composite tutorial on ansys for me as well:
hamed_riazi@yahoo.com
best regards

[AbdallaMallabdA]

Please Mr. Matt, could you mind send me a copy of the ANSYS Tutorial about Composites.

I'm a research assistant at Banha Faculty of Engineering - EGYPT.

My M.Sc. Research is on the studying of adhesive joining of composites and enhancing it by CNT.

I know it is difficult, as I'll do modeling on ANSYS and make it practical.

I'd be pleased if you send me this tutorial on my email: abdalla_19876@yahoo.com, as soon as possible.

Best Regards,
Abdalla M. Abdalla