Added c5g7 benchmark problem for MG mode

[nelsonag]

Forget the previous PR; I had some odd commits from 2013 apparently that made no sense.

[nelsonag]

Quick note: this isn't 100% ready for merge (and it makes no sense to do so anyways without #494 of openmc being merged first), I am also adding the c5g7 extended 3d benchmark and I want to cleanup the xml input files a bit. I put this PR in a little early so a multi-group model would be available for the reviewers of #494.

[nelsonag]

Now this is ready.

[samuelshaner]

What do you think about just having the Python input files instead of xml to minimize code reuse? I've created some OpenMC Python input files for C5G7 here:

https://github.com/samuelshaner/openmc/tree/mg/examples/python/c5g7

[nelsonag]

I see no problem with either way
On Dec 31, 2015 3:45 PM, "sam" notifications@github.com wrote:

What do you think about just having the Python input files instead of xml
to minimize code reuse? I've created some OpenMC Python input files for
C5G7 here:

https://github.com/samuelshaner/openmc/tree/mg/examples/python/c5g7

—
Reply to this email directly or view it on GitHub
#31 (comment).

[nelsonag]

@paulromano is there a maintainer-preference for input format (i.e., python vs xml)?

[paulromano]

Using the Python API would be my preference.

[nelsonag]

I'm getting there, albeit slowly. Comparing my results with those from @samuelshaner first to see if there any differences of note.

[nelsonag]

Ok, this is ready to go. I used the python api input format from @samuelshaner, which I hope he doesnt mind.

Here is a summary of results:

Case	Reference Sol'n	OpenMC Sol'n	Error [pcm]
UO2 Pin	1.32519(11) (From MCNP, not in official standard)	1.32555(5)	27
2D	1.18655(8)	1.18662(11)	-6
3D	1.18381(8)	1.18368(6)	11
3D unrodded	1.14308(6)	1.14311(6)	-3
3D rodded-A	1.12806(6)	1.12811(6)	-4
3D rodded-B	1.07777(6)	1.07767(6)	9

The diffrences are on the same order of, or significantly less than, the spread in eigenvalues published with the benchmarks for other MC MG codes (i.e., see table 3 of the original C5G7 spec).

The UO2 Pin problem is not in the official standard, but I've seen it in a few papers as the 'first problem' done so thought it worthwhile. The reference solution above is from MCNP (MG) done by the MPACT guys (this # was from their validation files). I cant speak to how right it is or is not.

[samuelshaner]

@nelsonag - the results look great! Glad to see you used the python input files.

[samuelshaner]

Can you remove the extra two spaces before rows 2+ in this lattice? This is probably an issue that I had in my inputs.

[samuelshaner]

Just tested all the inputs and everything checks out. I think this is ready to merge once the spacings is updated in lattices.py

[nelsonag]

Great, thanks for the quick review. I will get to the spacing tonight.

On Tue, Mar 22, 2016 at 9:43 AM, sam notifications@github.com wrote:

Just tested all the inputs and everything checks out. I think this is
ready to merge once the spacings is updated in lattices.py

—
You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub
#31 (comment)

[nelsonag]

Fixed,

[paulromano]

Thanks @nelsonag (and @samuelshaner for reviewing)!