Order of Lattice Indices

User Support
1

Hello,

I’ve searched but found no answer.

I’m filling a lattice and my Y indices are inverted or flipped with respect to what I would expect.

In MCNP on the lattice line I can change the order of the surfaces to ‘switch’ the direction of the indexing. And I can do this for only Y.

Is there a way I can do this in OpenMC?
My only other solution is to go back to my lattice and use a script to re-order it.
Rotations won’t work as the geometry is now ‘reflected’.

For reference:
https://mcnp.lanl.gov/pdf_files/la-ur-09-0380.pdf
Figure 5-5 PDF page 74

I am using OpenMC version 0.13.0-dev.
I can not share the input.

Regards,
Perry

2

hhhmmmm, well I think I just answered my own question.

I tested putting the Y pitch as negative, and it seems to have worked!
I did not expect that that would work.

Perry

3

Well, it seems this cat sung too soon :crying_cat_face:

The openmc-plotter shows exactly the geometry I want with my lattice pitch of “1.0 -1.0 1.0” and the appropriate transformations.
But the execution of openmc gives numerous “had a negative distance to a lattice boundary” lost particles,
which I am pretty sure are due to my negative pitch.

So the original question stands, unless any other ideas for my negative pitch.

Perry

4

Hey @yrrepy. Yes, in OpenMC when you specify the universes for a rectangular lattice, the y-indices are reversed (goes from higher y coordinates to lower y coordinates). This makes it such that when you “look” at your code, e.g.:

lattice.universes = [
    [u1, u1, u1],
    [u1, u1, u1],
    [u2, u2, u2]
]

it corresponds visually to what you would see if you took a slice through the geometry (i.e., what you’ll see in the plotter). If you want to reverse the y direction, the best thing I can suggest is to use numpy indexing capabilities. So, set the universes using a numpy array:

lattice.universes = np.array([
    [u2, u2, u2].
    [u1, u1, u1],
    [u1, u1, u1],
])

and then

# reverse y direction
lattices.universes[:] = lattice.universes[::-1, :]

Our lattices were never designed to work with negative pitches, although perhaps it wouldn’t be too difficult for us to make some fixes in the code such that it would.

5

Hi Paul, thanks.

Yes this is pretty much what I did.
I used a Python Pandas/Numpy script to take my lattice and reverse the ordering of the Y indices.

Cheers,
Perry