Functional expansion tally-based Shannon-Fisher complexity metric for fission source convergence analysis in Serpent 2 Monte Carlo code

This study presents a new fission source convergence metric, describing an information-theoretic spreading based on Shannon-Fisher statistical complexity measure. The metric is derived from a functional expansion tally-based (FETs) formulation by decomposing the fission source distribution into its spatial moments. The metric analyses whether those moments (and their correspondent expansion coefficients) converge upon an established stopping criterion. The Shannon-Fisher complexity metric provides global and local information from the Shannon entropy and the Fisher information measures. The evaluation of the new metric is performed on a single-assembly PWR test-cased described from the BEAVRS benchmark specifications. The implementation in Serpent 2.1.32 represents a stationary diagnosis metric, alternative to Shannon entropy, but potentially, a stopping criterion between inactive and active cycles