Simulating a large distribution of hard spheres

Geometry and Materials
1

Dear all,

I implemented a pipeline to simulate a tomography measurement of objects consisting of volumes containing very large amounts of hard spheres of equal dimensions, and given packing factor. An input file is generated per angle using a python routine, and mgdraw for scoring. I tested it for a relatively small population of spheres (around 14000), and all worked as expected. However, scaling to tighter packing factors (around 60000 spheres), I ran into the following error:
“STOP Flkcgi: more than MXEXEL elements in the zone”

I guess this means that I have reached the maximum amount of elements that are allowed, supposingly in the geometry definition (the spheres are indeed grouped in a single region). Is there a way to fix this issue?

Thank you very much in advance!

Riccardo

2

Dear Riccardo,
without seeing your input, I can only assume you have all your spheres in one Region. If you split that Region into multiple smaller ones, containing fewer spheres, you should be below the MXEXEL limit of 100’000.

Cheers,
David

3

Dear David,
thank you for your reply. I tried to split them into multiple regions, but still getting the same error. Attached are the input and log for one specific angle.

Cheers,

Riccardo
input_tomo_TRISO_0.0.inp (3.7 MB)
input_tomo_TRISO_0.0001.log (52 Bytes)

4

Dear Riccardo,

could you try to split the MATRIX region as well?

Also, does it make sense to have so many very small spheres? Wouldn’t be enough to calculate a homogeneous mixture material?

Cheers,
David

1 Like