I am modelling a single fuel assembly of 17 by 17 fuel rods composed of fuel, clad and water. The whole model has dimensions (26.50 x 26.50 x 500) cm. I used vacuum boundary conditions around the entire model. When i calculate the Shannon Entropy it shows the following behavior:
It does not look converged to me, even after 500 cicles. Shouldn’t it converge quickly for such a simple model?
When i switch the boundary conditions to reflective the Shannon Entropy appears to converge almost immediately. Does the boundary conditions affect the source convergence?
I ran this calculation with 50 batches (30 inactive), 10 generations per batch and 200000 particles. I used the following mesh to calculate the Shannon Entropy:
The slow convergence probably has to do with the z dimension (500 cm). How is your source defined? If you are using a point source, it may take many generations for particles to propagate all the way to the top and bottom of the fuel assembly. You should make sure that your starting source guess covers all fissionable regions. The easiest thing to do is to use openmc.stats.Box and have the lower-left and upper-right coordinates cover your assembly.
I see – I missed your note in the first message about using vacuum boundary conditions on all sides. Usually a single assembly model would be simulated with reflective (or periodic) boundary conditions in the x- and y- directions, and vacuum boundaries in the z direction. In the case where you have vacuum boundary conditions all around, it will indeed take a while for the spatial shape of the flux in the radial and axial directions to resolve.
