Best FEA software

[mrmartins]

Hi,

I am looking for a software to run impact simulations over welded metallic structures subjected to large plastic deformations, due to the impact. Any opinion about the best software, or the best relation qualty/cost?

Thanks in advance.

 

[cloche]

LS-DYNA from ANSYS, or Algor Multiphysics, or ABAQUS... Each of them has pros and cons, of course...
In particular, I've seen Algor Multiphysics Mechanical Event Simulation and it really impressed me... Worths a look!

 

[gwolf]

I think that once you decide to do impact and plasticity you are already into expensive software, so cost should be a less important criterion. I don't know of any cheap software that does these things well if at all.

This leaves you with ABAQUS/Explicit, DYNA in all it's forms, and ANSYS.

DYNA is a bit of public domain code like NASTRAN which has been packaged up by various companies and sold under different DYNA-related names. I recomend staying clear of it because it's buggy and awkward to use. Interfaces from pre-processors aren't great either. Some of the main software writers moved to HKS about 10-12 years ago and wrote ABAQUS/Explicit. I have used two types of DYNA and ABAQUS/Explicit and ABAQUS is far superior in every respect.

ANSYS- I don't like the user interface and the job control for the type of multi-step analyses that you want is poorly defined and difficult to understand.

However: ABAQUS/CAE which you would probably use with ABAQUS/Explicit is not a particularly full-featured package and if you bought ABAQUS/Explicit you would probably want another pre-processor like PATRAN, HYPERMESH, or CATIA and all this could get really expensive.

I suggest you invite the software salesmen for ANSYS and ABAQUS in and ask for a free trial version for 1-3 months. They may insist on sending the evaluator on weeks training course which you might have to pay for - this is not unreasonable.

 

[cloche]

As you want to perform drop-test simulation, implicit solvers are definitely out because they would hang on "unconstrained model". So comparing ANSYS "not-LS-DYNA" with ABAQUS/Explicit is misleading. I personally never did any, but I saw very sophisticated explicit analyses made with ANSYS LS-DYNA which were benchmarked against experimental data with great success. I use ANSYS every day or so, and yes I agree that it is VERY difficult to understand. I don't agree when Gwolf says that it gives poor control on the problem's parameters: reversely, I find it to be perhaps the "less automatic" and the most "user-controllable" FEA on the market (once again, it will take a long time and a lot of effort to understand the "internal" parameters and options).
Algor MES is perhaps the less-known solver to combine direct kineto-dynamic simulation with implicit structural formulation altogether at the same time in the same analysis. I evaluated it a year ago and I believe it would be a nice possibility for drop-test. MES' solver is a sort of flex-body kineto-dynamic simulator, and takes into account any form of non-linearity. Algor Full Multiphysics with MES and CFD was priced to me 22000 USD last year.
On the other hand, I seem to know that ABAQUS is perhaps the only solver that can handle fracture mechanics "directly" (I mean, without getting mad with birth/death of elements and coupling equations, as ANSYS would require), so from what I know it's the first-choice solution if you want to examine complete failure of your component due to impact (but treat this sentence with caution: I'm not sure...).

 

[Drej]

Here's my two penneth worth.

I'm a big ANSYS fan, and I've been using it a long time now, so when most people ask me for recommendations on FE software I always suggest ANSYS. However, I have to echo the sentiments of gwolf above regarding the initial steepness of the learning curve required and the userbility of the GUI (very steep and not that friendly, respectively - you get used to it eventually). That said, ANSYS would still be my general FE tool of choice. When you start talking high speed simulations, the ANSYS implicit code is probably not your choice (depending on your definition of high speed - though you might be able to do it in ANSYS. Plasticity is not a problem (at least not in metals here) but that might be another issue for you.). I would consider only two main packages: ABAQUS/Explicit and DYNA in one of its forms. If you've used ANSYS, I would suggest you seriously consider using ANSYS LS-DYNA (it's really a software made by boffins for boffins). It is seriously very good, if a little cumbersome and awkward to use per se. It's also relatively cheap (depnedent on the interface you choose to use with it). Compare that to ABAQUS/Explicit, which I've also used. It is an amazing code, very forgiving, easy to pick up (with excellent manuals) and much, much easier to use than DYNA (for the reasons outlined above). The only problem you may find is cost. The ABAQUS code is bundled (at least it was last time I looked) with /Explicit /Standard and CAE etc. You get all of these codes for the single licence fee. Ask yourself do you need access to all of these ABQUS features? Ask yourself is your geometry and mesh connectivity important (i.e. do you need good CAD/FE integration? You may need to spend further on an appropriate pre/post-processor for DYNA and ABAQUS). Do you consider learning costs important? Etc.

Cheers,

-- drej --


------------
See faq569-1083 for details on how to make best use of Eng-Tips.com

 

[GBor]

cloche,

Nice job on discussing Algor's MES capabilities. I have used Algor since well before it became a Windows-based product. The MES capability works well for drop tests and low strain-rate impact analyses. I didn't see the words "drop test" in the original post, but rather "impact simulations" and "large plastic deformation". Algor does both of these well.

High strain rate is another world. I have no personal experience, but I think LS-DYNA is recognized as the leader in this area. It comes with a high price tag, but it's worth the speed and accuracy from what I understand.

For low-strain rate, large deflection, impact simulations, Cosmos, Algor, and NENastran each have a product in the market. I'm not that familiar with Cosmos or NENastran's capbilities, but any of these three would likely be more cost efficient that LS-DYNA, and I can vouch for Algor's accuracy.

 

[ThomasH]

First let me say that I have very limited experience with explicit dynamics.

However I read recently that there is a new solution sequence coming shortly i MSC.Nastran called SOL700. This is LS-DYNA under a MSC.Nastran bulk data shell. From what I've heard LS-DYNA is better than most in the explicit area and that configuration would mean that you can access a good pre- and postprocessor as well.

Just a thought.

Thomas

 

[GregLocock]

Why not save yourself a heap of grief and use what your competitors do? Then you can recruit their analysts as well.



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[GBor]

Or be recruited by them?

 

[GregLocock]

Yup. It's a free market.

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.