Strange current behaviour at boundary

[lorenzoloi]

Hello everyone,

I’m currently performing some simple calculations in multi-group mode, aiming to extract group flux and current.

I’m considering a 2D square made with only one material and vacuum boundary condition, with side 100 cm and a regular mesh 40x40. For the current score, as the documentation sugest, I’m using a MeshSurfaceFilter.

My question is related to the current value at the boundary: since vacuum BC is applied, it is expected a score of 0 particles incoming. However, the outgoing score (i.e., the number of particles that crosses the surface from the square to the vacuum) should not be zero.
For all the test I performed, the outgoing score at the boundary is either 0 or has magnitude 1e-8, making it negligible with respect to its neighbour cell.

It seems that the MeshSurfaceFilter do not catch the crossing toward the vacuum. What do you think about that?

[lorenzoloi]

Hello again,

I would like to share with you an update: I tried to run a similar case in 1D, adopting both MeshSurfaceFilter and SurfaceFilter for the current score.
In order to compare the scores with the same mesh size, I have divided the entire geometry in as many cells as the mesh size, so that (if I understood correctly) the two filters should behave in the same way (apart from the fact that MeshSurfaceFilter gives the partial currents at the boundaries, while SurfaceFilter gives the net value at the surface).

To have a comparison with an analitical solution, me and my colleagues also calculated the expected neutron current from a diffusion approximation.

The diffusion result seems to be in agreement with the SurfaceFilter one. Is there a bug in your opinion?