Source file format

User Support
1

Hello everyone!
I was using a binary source file type as the external source, but it seems that openmc does’t recognize in the latest 0.7.1 version. Dose that mean the source file must be in HDF5 format? If so, how can I convert the binary to HDF5? Can anybody guide me？I would appreciate your help.

2

Hi Shawn, I’ve attached a Python script which will convert a binary source file into an HDF5 source that can be used with OpenMC 0.7.1+. Let me know if you run into any problems with it.

Best,
Paul

convert_source.py (1.12 KB)

3

Hi,Paul. Thank you for your kindly help. Here I’ve got anther question.In fixed source calculaton, the source_banks of different batches are totally the same(the source was sampled from given distribution and the space type is “point”). This really confused me .Is that used to eliminate correlation between realizations of random variables?
Regards,
Shawn

在 2016年4月9日星期六 UTC+8上午3:39:51，Paul Romano写道：

4

Hi Shawn,

I’m not sure I fully understand your question. By definition, a point source will result in every source site being born in the same spatial location. The angle and energy distributions are sampled separately so depending on what you chose for them, you should still see differences between batches (just not spatially though). As far as correlation between realizations of the random variables, that really only becomes an issue for k-eigenvalue calculations where the source sites for one generation are generated via fission sites from the previous generation. In that case, there is no easy way to get around the correlation.

Best,
Paul

5

Hi,Paul. Thank you for your reply. I may not expressing well in the question. The uvw and E in different batches are the same. I’ve attached my settings and geometry files and some output statepoint files. You might see what I did wrong.
Regards,
Shawn

在 2016年4月12日星期二 UTC+8下午8:28:09，Paul Romano写道：

statepoint.zip (2.33 MB)

input.zip (4.58 KB)

6

This was actually due to a bug. The source sites were supposed to be resampled at each generation but were not. This has now been fixed in the developmental branch for OpenMC and will be included in the next release.

Best regards,
Paul