Mystran a FEA solver for NASTRAN input decks is open soucre

[johanne...@gmail.com]

Hello Renaud,

I came to know that Mystran a FEM solver which is using the Nastran "language" and the Nastran output (f06 and op2) is open source. It may helps you to improve FeResPost.

You can find the source code (FORTRAN 90) under the following Link:

http://opensource.gsfc.nasa.gov/projects/mystran/index.php

Mystran is released under the terms and conditions of the NASA Open Source Agreement (NOSA) Version 1.1 or later.

Under http://www.mystran.com/  you can download a Windows and a Linux binary but it looks like that a license key is necessary to run it.

I hope you like it.

(I posted the same to the pynastran team)

Best regards, Johannes

[Renaud Sizaire]

Hello Johannes,

Thank you for your interest in FeResPost.

I will try to install Mystran to test FeResPost. I presume that the reading of BDF and OP2 files should work with Mystran as well as with MSC/Nastran.

Regards,

Renaud.

--
You received this message because you are subscribed to the Google Groups "FeResPost" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ferespost+...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Steven Doyle]

Renaud,

I posted this in the pyNastran discussion page and I'll post it here as well.

I have some concerns about the license in terms of integrating it (it's a NOSA license and there is a term the FSF takes issue with that seems to be incompatible with LGPL).  I'm not sure I agree with the FSF, but I'd like to resolve it.  I've sent an email to the point of contact at NASA.  

The original term per http://opensource.org/licenses/NASA-1.3

> Each Contributor represents that it's Modification is believed to be the Contributor's original creation and does not violate any existing agreements, regulations, statures or rules, and further that Contributor has sufficient rights to grant the rights conveyed by this Agreement.

The complaint is per http://www.gnu.org/licenses/license-list.en.html

> The NASA Open Source Agreement, version 1.3, is not a free software license because it includes a provision requiring changes to be your "original creation".  Free software development depends on combining code form third parties, and the NASA license doesn't permit this.

Regardless, there is probably some value in just getting MYSTRAN working and calling it as an executable (rather than incorporating it into the source).  I took a look at it and at a minimum, it supports statics, modes, buckling, and nonlinear statics.  It's old though (there are Pentium 2 flags).  It looks like it was last worked on in 2012 though.  I'm pretty terrible at compiling Fortran, so if anybody knows how to convert Lahey Fortran to gfortran and fix linking errors, I'd really appreciate the help.  I've added Mystran to https://github.com/SteveDoyle2/mystran with the same NOSA license.

MYSTRAN is definitely following a non-standard version of Nastran.  The BDF case control has different rules, there are some new cards, there is support for some long since retired cards, and the PARAM card has a max of 4 value fields instead of a max of 2 (e.g. "PARAM,AUTOSPC,Y,,0.001").  The F06 is also very, very different.  Still, if there was a working build, it wouldn't be too hard to bring it into somewhat better compliance (at least the BDF).

Steve Doyle

[Renaud Sizaire]

Hello Steven,

Thank you for this information.

Actually, I have in mind to test the reading of BDF and OP2 of Mystran. It should not be too difficult to adapt FeResPost NastranDb class and make it work. This does not involve the use of Mystan source code.

Regards,

Renaud.

[Steven Doyle]

When you decide on how you're going to do integrate it and find the gotchas, can you let me know?

It somewhat concerns me because pyNastran is beginning to diverge in how it handles the BDF and OP2 for NX and MSC and it's complicated enough as it is.  There are some subtly different rules between MSC and NX such as:

1.  BDF node/element/others? IDs in NX can overbound the 8-character field

2.  BDF SET3 cards have different rules (e.g. single-lined for NX vs. multi-lined for MSC)

3.  OP2 tables have different table codes (e.g. the MPCFORCES table).

4.  OP2 force/stress/strain element types have different element type IDs

4.  NX Nastran will give you the version (e.g. NX8.5), but MSC returns XXXXXXXX.  This depends on the PARAM,POST option you use, so it's not reliable, but it is a way to auto-detect the version.

PyNastran now has support for the following at the top of the file to avoid various problems, which would help in supporting Mystran.  It probably makes sense to come up with a common header.  The current header is:

$ pyNastran: encoding=latin1

$ pyNastran: punch=True

$ pyNastran: version=msc

I'd like to avoid changing the well established rules to support MYSTRAN. If it's going to be supported, I'd like to make it very isolated from the rest of the code (e.g. have a BDF and MystranBDF class) and just overwrite some cards.

I'd also like to not support it at all until we have a build of it that doesn't expire after a month.

Thanks,

Steve

[Renaud Sizaire]

Hello Steven,

I confess I never tried to read NX Nastran BDF or OP2 files. I think it should be possible to update FeResPost NastranDb class so that it adapts to a specific version of Nastran. For example, it might be done by adding an optional parameter when creating a NastranDb object.

I may be easier to adapt FeResPost to different versions of Nastran because it is devoted to post-processing only. One does not try to manipulate the FEM entities. If a new element is defined in NX Nastran or an old element behaves differently than with MSC, I can easily add new elements in FeResPost library. This is true for BDF as well as OP2 files.

Regards,

Renaud.

[Steven Doyle]

I suggest you don't support MYSTRAN.  I found one better.

NASTRAN Cosmic has been open sourced.  https://github.com/nasa/NASTRAN

[Renaud Sizaire]

Hi Steven,

Thank you for the information. Actually, I think I will try to support as many versions as possible.

Regards,

Renaud.