FEA Sofware Reccomendation 3

[savis]

Hello all.

I am a Mechanical Design Engineer with a company that manufacturers thermoforming equipment. We are currently looking at purchasing an FEA package to analyze several parts on our thermoforming equipment. Solidworks is our 3D design package. We plan on training for whatever package we choose.

The two packages we are looking at are COSMOS (can be purchased thru local Solidworks rep. and ALGOR.

A stripped down version of COSMOS is included with Solidworks.

Anyhow, are their any words of advise to assist us in choosing the correct software.

Knee-deep in 12" of rain over the past 7+ days.

savis

 

[bckemp]

To KTMer,

Am interested in your investigation. I have similar requirements for robust electronics packaging and am interested in your experience with Algor?

Am looking for a package that can do SS and transient heat transfer, static stress, and natural frequency.

 

[savis]

To bckemp,

We have completed our search and it looks like we are going with ANSYS. Mainly due to their local reps customer support (supported by Engineers who actually use the software - novel concept). They actually confirmed an analysis for us at no cost as a demo!!

Savis

 

[DandAltd]

I've just scanned the replies above and have the following comments with regard to IDEAS. Like FSI I too use Ideas as a FEA package and moved from Patran to Ideas because Ideas IS much simpler to use than Patran whilst maintaning a very powerful core FEA capability. The Ideas solid modeller IS very, very easy to learn (I could teach you the basics in a couple of minutes) and the intergration with Solid Modelling and FEA is second to none. It also has CFD (although they call it ESC) as an option (if you can afford the license) which I also use. It's not the cheapest option but in my opintion its worth the extra money for the full CAD intergration/FEA capability. In its favour with regard to cost is that you pay for a perpetual license at the start and only support thereafter, unlike Patran which has to be purchased each year.

 

[GregLocock]

I find it baffling in these threads that 9 times out of 10 people recommend the software they use themselves.

FWIW the company I work for probably has the biggest userbase of IDEAS in the world and we do not use its FEA capabilities for anything but the simplest of jobs. I can't believe a Solidworks shop would buy an IDEAS seat just to run IDEAS FEA.

Can the original poster specify the sort of analysis he needs to do, whether he needs assemblies, is version control important, does he want an automesher, etc etc?



Cheers

Greg Locock

 

[brad]

Greg,
Amen.
I have used most of the major FEA softwares, and have a fairly strong background with the MBD softwares. I have my favorites, but those are often out of the price range of smaller users.

To answer the original question: "What is the correct software?" is very user-, usage-, and budget-dependent. You are better off talking to people of a similar stripe/background than asking a bunch of anonymous people to spout off.

Sorry to vent, but these "what is the best" questions almost always devolve into dogma and personal biases, rather than a useful discussion.

Brad

 

[analyzer]

How does NISA compare to the rest of the packages in this thread?

 

[EnglishMuffin]

I've used NISA a fair bit. It has a good multi-physics capability and a reliable solver, but a terrible pre-processor, (or at least it did when I last used it). Maybe it has improved some.

 

[jetmaker]

NISA... geeseee analyser.. that was the first package that I used over 10 years ago... thought it had gone away. Who makes it now???

 

[EnglishMuffin]

EMRC, same as always. They are Indian guys.

 

[jagad5]

I would rate NISA as a capable tool but I would not consider buying today. I define three groups of tools:

First Tier (this is the only group where I list every member): Abaqus, Ansys, NASTRAN - these are the best commerically FEA tools available. I would not hesitate to use any one of these to solver any problem. That does not mean each is great at everything. Abaqus is THE tool of choice for any kind of non-linear problem. They can handle plasticity, non-linear elasticity, friction, huge deflection. NASTRAN has traditionally been the premier tool of choice for vibration problems. ANSYS is the one tool to buy if you can only buy one and must be able to solve every type of problem. I believe ANSYS is also the current revenue leader among FEA programs. (It's hard to tell since most of the others are sold by CAD companies, mixing the picture)

Second Tier (et.al.): I-DEAS, Mechanica - these tools are accurate for certain types of analysis. Mechanica provides excellent functionality for linear elastic problems. I-DEAS is an excellent pre/post processor and has a decent solver. Niether of these tools should be considered an equal of any of the three in the first group.

Third Tier (very large group): I won't name names but the tools in this list function (some even function well) but none can compete when you must have good results. Many on this list should only be used to design simple parts with simple loads, that have simple consequences if failure occurs. I dismiss many of of the programs on this list for their offensive sales practices alone. Think of any frequently advertised/editorialized/popularized FEA program and it probably belongs on this list. ANSYS, ABAQUS and NASTRAN don't have time to answer the phone to answer questions for a review article, because they're too busy figuring out better ways to solve hard problems (rather than spending time figuring out better ways to lure more butts into their seats.)

These are all just my humble opinions. Never buy a package until you have solved at least one of your problems to your satisfaction.

Merry Christmas. Feliz Navidad. Frohliche Weinachten.


Doug

 

[jetmaker]

jagad5 brings up a really good point. You should never buy any FEA package without first evaluating it yourself in-house. Any company serious about selling you a license should be willing to let you do a monthlong evaluation. At one company, we were benchmarking ProMechanica against AFEA. When importing complex geometries from CATIA databases, we encountered unexpected problems with ProMechanica. Something that the manufacturer said would not happen. Eventually the problems were worked out, and final results were found to be reasonably similar for the 2 packages. The difference was that Mechanica ran considerably faster. Therefore, this program was adopted to run alongside our AFEA machine as a system to run the preliminary anaylsis. Final and non-linear analyses were still relegated to AFEA. This approach allowed the company to save money in both time and cheaper licensing costs.

Best of luck in your decisions.
jetmaker

 

[EnglishMuffin]

You know, its a pity there isn't an industrial equivalent of "Consumer Reports", who could provide unbiased comparisons of software and industrial components of all kinds - there might be some surprises. Magazines are not fully trustworthy. But of course it will never happen.

 

[jagad5]

There is an independent organization called NAFEMS which creates "standard" problems with known answers that can be used to test FEA packages. I believe their website is

http://www.NAFEMS.org.

Many of the codes I like have tested their programs using these test cases. Its not quite Consumer Reports, but it does provide a means of making direct comparisons between different packages.

Doug