0

there is a source code for Natran-95 on Github. I need to install it on Window 10

but searching on the web there is no clear tutorial on installation. while I click on the nastran.exe on this package in the hyperlink to GitHub, I get windows error the file not found in directory.

the specific reason to use nastran is to generate a superelement from a PSHELL model. it seems that MYSTRAN 12.0 has added sparse solver that some of its subroutines do the reduction of sparse matrix, but I am not sure if it also saves it or only uses for solving, but uses the same craig bampton method

could anyone tell me if there is a compiling step missing ? just if you know how to install it, and also use it, please write an answer? it will be a great source for internet

FabioSpaghetti
  • 260
  • 1
  • 11
  • The original NASA versions of Nastran (like the Github link you mentioned) are completely obsolete. Note that the user manual in that link is dated 1986. Don't waste your time trying to install and run a 35-year-old version of the software, unless you want to do it as a retrocomputing project (and in that case, ask the question again on retrocomputing.SE) – alephzero Feb 01 '21 at 14:04
  • @alephzero According to Wikipedia the last stable release was in 2014. – Eric S Feb 01 '21 at 14:22
  • @EricS I think you are confusing the closed-source commercial version (begun by MSC in about 1970 and still being developed) with the original NASA version. The NASA version only ever ran on IBM, Univac, and CDC mainframes, and was originally supplied on magnetic tape which included the installers and machine-dependent code. It was never structured as an open-source project which users could build completely for themselves. The original version was put into the public domain as a legal requirement of NASA's US government funding. It is software archeology, not a useful modern FEA application. – alephzero Feb 01 '21 at 15:27
  • Actually, the Wikipedia page on Nastran might be the clue to answering the OP's question: "OCF Nastran: Both source and binary copies of Nastran are available from the Open Channel Foundation for an annual license fee. This was part of the NASA COSMIC Collection distribution from the National Technology Transfer Center. It was published in June 2015 on GitHub.[citation needed]" The 2015 date is consistent with the OP's GitHub link. – alephzero Feb 01 '21 at 17:15
  • @alephzero I really think it is open source and doesn't require a license based on the open source agreement document on Github. Doesn't mean it is easy to use or even that there is a Windows installer, but it does mean that the software is there and assuming you have a FORTRAN compiler useable. – Eric S Feb 01 '21 at 17:26
  • If you Google "nastran95 windows" you get some useful hits like https://github.com/nasa/NASTRAN-95/issues/5 and https://www.eng-tips.com/viewthread.cfm?qid=433397. You may also want to look at Mystran which looks like it is Nastran compatible and more up to date: https://www.mystran.com, which is compiled for Windows and is available as an .exe file. – Eric S Feb 01 '21 at 17:29
  • Thank you guys , I don't care it is hard or easy to use, the only thing is I Do not want to pay for it and apparently this is the only version which is open source. I don't have fortran compiler and don't know how to use it and even if I know, I don't know how to compile this one while there is an exe file already in the package – FabioSpaghetti Feb 01 '21 at 17:39
  • I noticed that there is a software called freecad which they have linked it to this Nastran 95 as a pre processor and another called hfcNastran is used for postprocessing but no clue how to compile nastran and connect them – FabioSpaghetti Feb 01 '21 at 17:42
  • Are you required to use NASTRAN? If not Wikipedia has a table listing a lot of FEM options. https://en.wikipedia.org/wiki/List_of_finite_element_software_packages – Eric S Feb 01 '21 at 18:08
  • yes, only and only Nastran, cause I need to generate a superelement from shell elements and Nastran has its own way of assembling the stiffness and mass matrices – FabioSpaghetti Feb 01 '21 at 18:23
  • The main question is not answered yet, there is no clear tutorial of full installation – FabioSpaghetti Feb 17 '21 at 07:05

2 Answers2

1

Use this fork: https://github.com/AeroDME/NASTRAN-95

Use gFortan and cmake to build it on windows 10.

Cean
  • 11
  • 1
  • Why isn't the built files uploaded there ? why should every user use gFortran and cmake ? I remember I tried it once and got lost in the middle – FabioSpaghetti Mar 08 '21 at 21:47
0

I hesitate to answer this as I have no personal experience with the software, but you should investigate MYSTRAN which is a free NASTRAN clone. It is available already compiled for Windows. From the about section of its home page:

MYSTRAN is a general purpose finite element analysis computer program for structures that can be modeled as linear (i.e. displacements, forces and stresses proportional to applied load). MYSTRAN is an acronym for "My Structural Analysis", to indicate it's usefulness in solving a wide variety of finite element analysis problems. For anyone familiar with the popular NASTRAN computer program developed by NASA (National Aeronautics and Space Administration) in the 1970's and popularized in several commercial versions since, the input to MYSTRAN will look quite familiar. Indeed, many structural analyses modeled for execution in NASTRAN will execute in MYSTRAN with little, or no, modification. MYSTRAN, however, is not NASTRAN. All of the finite element processing to obtain the global stiffness matrix (including the finite element matrix generation routines themselves), the reduction of the stiffness matrix to the solution set, as well as all of the input/output routines are written in independent, modern, Fortran 90/95 code.

The author was a member of the original NASTRAN development team and the software is free and open source released under the NASA Open Source Software Agreement. If you want to use the actual NASTRAN95 code, I'm pretty sure you'll need to have a FORTRAN compiler and compile from source.

I do recommend you consider alephzero's suggestion to pursue more modern options.

Eric S
  • 880
  • 7
  • 14
  • seems nastran 95 is the most developed one between them https://www.mystran.com/forums/showthread.php?tid=4 – FabioSpaghetti Feb 01 '21 at 18:09
  • @FabioSpaghetti Well, do you need the more advanced features? If you are unable to compile from source you might not be able to use NASTRAN95. Is there a specific reason you are required to NASTRAN95? – Eric S Feb 01 '21 at 19:09
  • yes the specific reason is to generate a superelement from a PSHELL model. it seems that MYSTRAN 12.0 has added sparse solver that some of its subroutines do the reduction of sparse matrix, but I am not sure if it also saves it or only uses for solving, but uses the same craig bampton method – FabioSpaghetti Feb 01 '21 at 19:15
  • @FabioSpaghetti Okay, you really should have added that info to your question. – Eric S Feb 01 '21 at 20:33
  • @FabioSpaghetti You seem to be missing the point that compared with any *modern* FEA software (including *modern* commercial versions of Nastran) you are quibbling about whether Brand X or Brand Y is "better" when one of them is 30 years out of date and the other is 31. (FWIW, I was actually doing FE analysis, developing FE codes, and networking with the guys who *founded* companies like MSC-Nastran, Abaqus, LS-Dyna, etc, more than 30 years ago, so I *might* know what I'm talking about from first hand experience here.) – alephzero Feb 01 '21 at 23:07
  • I'm only interested in the one that can output the super element stiffness mass and damping matrix – FabioSpaghetti Feb 02 '21 at 06:50