[SciPy-dev] Nastran generated sparse matrices was Re: Eigensolver

Nils Wagner nwagner@iam.uni-stuttgart...
Wed Apr 23 16:33:45 CDT 2008


On Wed, 23 Apr 2008 22:51:02 +0200
  "Stéfan van der Walt" <stefan@sun.ac.za> wrote:
> 2008/4/23 Nils Wagner <nwagner@iam.uni-stuttgart.de>:
>>  Who can help me writing an "interface" for these C
>>  routines ?
> 
> Is there a bounty involved?  

What do you keep in mind ?

I think many people would
>help you, given
> a bit of incentive.
> 

Well, Nastran is an industry-leading, linear and nonlinear
finite element analysis solver.

Therefore it would be nice if system matrices generated by 
NASTRAN [1] could be imported by SciPy.

[1] 
http://www.mscsoftware.com/assets/MSC_DS_MSCNastran.pdf

There are other finite-element packages, e.g. PERMAS [2] 
which  supports the output of system matrices in a 
user-friendly way :-)
You can use

K = coo_matrix((vals,(rows,cols)))

[2] http://www.intes.de

Abaqus 6.7 supports the output of system matrices, too.

I am not familiar with Ansys [3] but I think there exists 
a way to export system matrices as well.

[3] http://www.imtek.uni-freiburg.de/simulation/mor4ansys/

Nils



More information about the Scipy-dev mailing list