FEA which one to choose

I work for a manufacturer of automotive interior hinges. We recently benchmarked ANSYS, MARC, and NEiNASTRAN to replace our current MSC.NASTRAN (which couldn't handle our contact and large rotation well). All performed well and are worth a look.

COSMOS didn't make the first cut.

How does Algor stand up to this pack? We have the stand alone version and we're thinking about upgrading to the SW version (better interface and I think the data is stored in the SW files)

We didn't look at ALGOR. The consultant we work with didn't include it in his list of candidates, so it didn't come up.

Initially, we were looking at replacing MSC.NASTRAN with MSC.MARC, as recommended by our MSC rep (!). We wanted to benchmark at least two others, and NEiNASTRAN and ANSYS were the ones we chose with the help of a consultant we have on retainer for training and mentoring (and knows our work very well).

CosmosWorks and DesignStar are still works in progress for which you will find limitations in how far you can go. The older and more difficult to use Cosmos/M can do it all, but has an older interface.

I have both DesignStar and NE/Nastran available to me. ANSYS is also a mature product.

There are a couple things about CosmosWorks that you need to strongly consider.

1. When you run a long and large analysis:

a. It ties up SW for the duration b. All the memory that SW uses is not available for the FEA program to use.

1. If documentation of your analysis is important to you then you shouldn't change geometry in SW after making the run. This is both a strength and a weakness. It is a strength in that you can quickly search for the most efficient structure. It is a weakness when you have to document what you did in order to defend yourself when product X fails. Since SW files are always changing in size how can you prove the file you are looking at is the one you did the original analysis in?

A program like DesignStar can also link to actively changing geometry, but provides a standalone archivable file for documentation. And the solver can be run without the GUI for maximum memory usage. It can also run nicely along with SW by setting its priority down in task manager. I have run 60 hour analyses with DStar while using SW.

Keith Streich wrote:

I just ran across this link.

I recommend NEiWorks which is a Nastran pre and post processor and solver inside of SolidWorks. NEiNastran supports surface to surface contact. See their website for demos