Anybody enjoying ALG/Nastran?

I haven't used ALG/Nastran, but I have used the Nastran translator from Algor. I have run a FEMPRO-built model (Algor's pre- and post-processor) through NEiNastran and MSC Nastran. It has been about 2-1/2 years since I did this, but the results were comparable on all three platforms. The model wasn't terribly complex, but it was a little more than a flat plate. It was a beam-plate model with some non-rectangular geometry.

As for the "Nastran" term in ALG/Nastran, it raises the question of what "Nastran" really is anymore. MSC is clearly based on the original NASA code. Cosmic Nastran, may it rest in peace, would likely also be considered the "true" Nastran based on the NASA code with nearly 40 years of development. Are newer codes still using the 1970 solver? If not, are they really Nastran? If they are, is it still efficient? If they're not, does the "Nastran" label allow them some illegitimate confidence?

I'm not a developer, so smarter people than I will have to respond, but I think the term "Nastran" is over-rated and I hope engineers are validating the software and their use of it. If utilized properly, NEiNastran, Algor, and others, I'm sure, are excellent software packages. They all have their bugs that must be worked out.

My 2 cents

This issue is addressed also on the Algor forum

NASTRAN file translation
thread810-129702

We used to use Algor with ALG/Nastran and have since switched for several reasons, this product being one of them. If you do any serious analysis and require true Nastran compatibility do not buy this product. What Algor is selling is their solver with a Nastran translator to write out a bulk data file. They are pushing this as a viable Nastran solver. It is not. What we found is that for basic models with simple elements, loads, and boundary conditions, it translates out the correct bulk data. If you go the other way and read in a typical Nastran file it will miss or skip a lot of entries. This is very dangerous.

Just a few thoughts.

I would say that the "Nastran" name or label is not neccessarily overrated but very missused. ALG/Nastran and NEiNastran (as far as I know) use the Nastran bulkdata format and have their own solver, not based on NASA's code, to do the numbercrunching. MSC.Nastran and NX Nastran are based on the original code and a lot of development.

I thing that in some respects using the name and not the code is marketing and nothing else. It doesn't mean that the procuct is bad. But the sales department doesn't seem to think that it can sell "alone".

As I understand from Noran NEiNastran is based loosely on the original COSMIC version just like MSC and NX but was adapted for the PC in 1990 and many parts were replaced or completely updated with newer element formulations and Fortran 90 code. Much of MSC has also been replaced over the years. Our models are definitely complicated and are routinely compared to MSC since other parts of our company still use that brand. Any time there has been a difference we just refine the mesh and re-run in MSC to get the same result NEiNastran got with the coarser mesh. To call what Algor has done a Nastran would mean anyone with any FEA program that writes out bulk data can just repackage it as Nastran. I understand MSC is now adding all new features in Marc which is now under the guise of solution 600. I think it would be safe to say that MSC is still the defacto standard and that their development was very significant but much of it as of late was to appeal to the 10% of their customers that brings them 90% of their revenue and not to the majority. That is why we left them. A measure of how "Nastran" a product is has more to do with how they handle all the subtle nuances of Nastran as defined in the early NASA Theoretical Manuals. Things like transformations of skewed boundary conditions, how the element coordinate system is defined, how temperature dependent materials are treated, etc. are good indicators. What MSC has done for Nastran has not always been good and hence the law suit by the FTC. I personally think the enhancements from Noran have been outstanding and obviously MSC agrees as they have been copying them in their latest versions. Maybe the best solution is to ban all companies except the ones with the original 1970 COSMIC version from calling their product Nastran including MSC. The only problem is there would be no more Nastran products if you did that. The only true Nastran’s out there are NEi, MSC, NX, ME, and TMG versions. All the rest including ALG, VR, ABC, etc. are just codes repackaged for marketing purposes.

Frank,

Not to be argumentative, but I don't understand:

...no more Nastran products

Click to expand...

followed by

only true Nastran's

Click to expand...

Sounds like there really are no true Nastran's at this point, so what does marketing repackaging matter? If the answer is right, it's right. If it's wrong, we need to make sure as analysts that we are properly using the product. If we are and it's STILL wrong...then there is a serious problem.

I agree that NEiNastran is a great product and MSC is living off old glory. I haven't used NXNastran or any other flavor, and I'm not really concerned with what Algor has done with ALG/Nastran, but haven't all of these packages basically

[written] out bulk data...[and] repackage[d] it as Nastran

Click to expand...

?

I think what Frank is saying is that none of them are true Nastran products anymore. They all have evolved. Garland is absolutely correct that all that really matters are the results.