openmc-dev
Shannon Entropy for Random Ray
Similar to Monte Carlo, computation of Shannon Entropy is a useful way to determine if enough inactive batches have been run such that the source has fully developed and become a stationary distribution.
In Monte Carlo, the Shannon Entropy is calculated using the number of fission site particles that are sampled from each cell within an overlaid Cartesian mesh. However, with random ray, as we already have a well refined spatial mesh (due to the need for FSRs), it is likely sufficient to compute the Shannon Entropy using the source terms of each FSR directly. I.e., there is no need to define a secondary spatial mesh. Additionally, as the Shannon entropy can be trivially deduced from the source, it may make sense to always compute it rather than requiring a user input setting to be turned on.
Note that Shannon Entropy will also likely be useful when running in fixed source mode, as in random ray inactive batches are still needed to converge the scattering source.
