MSC Nastran 2014 *.f06 output significant digits

[timpap]

Hello to all,
I am using MSC Nastran 2014 for a linear static analysis. I would like to gain more accuracy in my output displacements than the 6E13.5 format that exists in the *.f06 files.
I have tried using the long field format in the BULK DATA entry for the geometry but that doesn't affect the output significant digits.
Is there a way to change that?
Thank you in advance!

PS: I forgot to mention that I would like to do that through the input file, so I guess I am looking for some kind of parameter to add.

 

[Zadrobu]

I don't know of any way to output more significant digits in a .f06 or .pch output file. I believe it's possible though to get 16 significant digits in an .OP4 output file.

But what are you going to use the displacement results for? Are you really sure that you need the displacements in that many significant digits?

 

[timpap]

Yes, i need to make a sensitivity analysis. Can i get the displacement results in the .op4 output file?

 

[Zadrobu]

It's possible to output displacement results to .op4 output file with a DMAP alter. Put the following statements in the Executive Control section of your Nastran runfile (i.e. put it at the very top of the file):

$$ Assign OP4 output file
ASSIGN output4='displacement_matrices.op4',unit=91,form=formatted
$
$$ Use linear static solver
SOL 101
$
$$ Print out G-set displacement matrices
COMPILE SUPER3 souin=mscsou,nolist,noref
MALTER'BOTTOM OF SUPERELEMENT DATA RECOVERY LOOP'
OUTPUT4 UG,,,,//0/91///15 $
ENDALTER
$
CEND

Nastran should then generate a file 'displacement_matrices.op4'. The formatting of OP4 files is not very human-friendly, but I think that you should be able to make sense of the numbers by comparing them with the results in your f06 file.

 

[dmapguru]

There is a small risk with the DMAP alters suggested by @Zadrobu that the order of the output printed to the op4 file will not be the external sequence. Matrix UG has its rows ordered in internal sequence, which for the vast majority of cases corresponds to the external (i.e. user defined) GRID point sequence. However, if one of the more rarely used solvers is activated (e.g. iterative or active column), there is an outside chance that the row order of UG will not be the external GRID point order. In this case, you will find the order of the displacement values in the op4 file does not match the order of values in the f06 output (standard f06 output is in external order).

You can mitigate against this by always transforming UG to external sequence before outputting it to op4, like this:

MATGEN EQEXINS/INTEXT/9//LUSETS $
MPYAD INTEXT,UG,/UGEXT/1 $
OUTPUT4 UGEXT,,,,//0/91///15 $

If the internal order matches external order, matrix INTEXT will be an identity. If not, its permutation terms will render UGEXT in external order.

A final note: just remember that the displacement output (f06 and op4) is in the global coordinate system. If you have local coordinate systems, and need the displacements to be transformed to the basic coordinate system, that is also possible with an additional call to the VECPLOT module.

 

[timpap]

Thank you both for your great answers! You really helped me! Thankfully I am only using the linear static solver so it seems like the grid points are in the same order as in my geometry file.
Just for further reference, could you recommend me any tutorials which can help me learn DMAP alters so I can control the output of the analyses better?
Thanks in advance

 

[Zadrobu]

Glad we could be of help. I'm still pretty much a novice to DMAP myself, but I've been trying recently to gain some more knowledge on it. Dmapguru's tip was a good one, I didn't know about the internal/external ordering yet.

These could be good places to start for learning more about DMAP:
[ul]
[li]DMAP Programmer's Guide[/li]
[li]Programming DMAP in MSC Nastran by Mark Robinson[/li]
[li]MSC Software DMAP forum[/li]
[li]Nastran DMAP Yahoo Group[/li]
[/ul]