Nastran: Core problem?

[Wall2]

Hi,
I'm doing a SOL 103 analysis of a somewhat large model, > 2 million nodes.

Everything went fine until this night when this was written to the .f04-file:

0:39:22 1873:43 2725.5G 1.0 66161.0 0.5 SEMR2B 43 MPYAD BEGN
METHOD 1 NT, STORAGE 3, NBR PASSES=****, EST. CPU=999999.0, I/O= 1717.4, TOTAL=999999.0
0:39:22 1873:43 2725.5G 0.0 66161.0 0.0 MPYAD BGN P=9999
MPYA1Y/ CORE PROBLEM21125433022106835498
METHOD 1 NT, STORAGE 1, NBR PASSES=****, EST. CPU=999999.0, I/O= 1717.4, TOTAL=999999.0
0:39:22 1873:43 2725.5G 0.0 66161.0 0.0 MPYAD BGN P=9999


Nastran is not beeing very specific here, does anyone know in more detail what CORE PROBLEM mmeans?

 

[ykantor1]

probably your model size is too big for Nastran current setting.
Do you use Patran to run Nastran , or run Nastran as stand alone?

look at Nastran installation manual, for the parameter which enablr big models.

good luck

 

[BlasMolero]

Dear Wall2,
A freq analysis of a model with two million nodes will require a lot of hard disc space, as well as RAM memory, check the *.F06 for information of the error, with NX NASTRAN the *.f06 file is the first place to look for the reason of any error of the NX NASTRAN solver, you may run out of disc space of not enought RAM memory was specefied.
Best regards,
Blas.

~~~~~~~~~~~~~~~~~~~~~~
Blas Molero Hidalgo
Ingeniero Industrial
Director

IBERISA
48011 BILBAO (SPAIN)
WEB:

http://www.iberisa.com

 

[Wall2]

Thanks for your replies and you might be right that the model was too big for my system, however, the Nastran calculation actually never stopped and a week later I had my results.

The calculation took 203 hours and maximum scratch space in use was 600 Gb.

The biggest lesson I have from this is: keep the node/element count to a minimum.