Thanks for all that info, it helped me a great deal. After talking to my VAR, it seems COSMOS/M can do non-linear surface-to-surface contact simulations, so i'll be exploring that soon, as well as maybe the next version of COSMOSWorks, which will apparently incorporate non-linear simulations. I've also been in contact with Ansys and they say it's no problem (i've seen ansys in action at the local university, they have a bunch of educational licenses). Seems like things are on track for me, thanks.
Andre Richard
> Thanks everyone, those are all good suggestions, but i'm wondering if
> > it can really help me in my case. Here's an image of the process: > >
> > http://24.230.198.245/Forming.jpg> >
> > The sheet gets pressed about 1/2" or so in the "Depth", then released,
> > the sheet advances about 2 inches, and is then pressed to the same
> > Depth again. I would like to be able to compare a typical situation we
> > do often, with variations in Depth, Thickness, Distance, Material,
> > even if for only one "hit" (i can only imagine what kind of processing
> > would be required to analyze the whole process of hitting it about 75
> > times). Note that the "Z" direction of my drawing would be uniform
> > (usually about a 3 foot wide strip is pressed at a time, for lengths
> > varying from 3 feet to 15 feet). It's sort of an "iterative" roll
> > forming process. I hope this helps in understanding my situation. >
> Andre,
>
> This is some pretty heavy duty FEA. You'll need non-linear materials and
> large deformation. If the punch moves very far, you'll probably need code
> that remeshes the plate. I suspect that you should hire a consultant to do
> it, as there will be a pretty steep learning curve for you to get up. If you
> need to do this a lot and want to have the capability in-house, then expect
> to pay quite a bit for the code, quite a bit for a really powerful computer
> to run it, and plan on quite a lot of time to run the analyses. And plan on
> quite a bit of time choosing the code you want.
>
> As far as codes go, Abaqus specializes in this type of analysis. Nastran,
> LS-Dyna, and Ansys are also possibilities. At the low-cost end of the
> spectrum, Algor is supposed to be able to handle these types of problems,
> but I don't know how well it really works.
>
>
> Jerry Steiger
> Tripod Data Systems
> "take the garbage out, dear"