Neutron clustering as a driver of Monte Carlo burn-up instability

• DOI: 10.1016/j.anucene.2019.106991
• Publication Date: 2019-08-22T17:59:50Z
• Authors: P. Cosgrove; E. Shwageraus; G.T. Parks
• Keywords: Depletion; Neutron transport; 0202 electrical engineering, electronic engineering, information engineering; 02 engineering and technology; Monte Carlo; Stability; Clustering

Coupled Monte Carlo neutronics and depletion problems have been noted to produce nonphysical power oscillations for large, spatially-decoupled problems. Previously these oscillations have been attributed to `numerical instability' - this work proposes that a prominent contributor to this phenomenon is neutron clustering during the Monte Carlo simulation, resulting in a poor estimate of the transport solution. This is demonstrated using insights from recent work on clustering and applying it to standard practice for Monte Carlo/depletion problems by performing simulations with the same number of histories - both total histories and only active histories - but different numbers of particles and cycles. The results demonstrate that neutron clustering appears to trigger instabilities in the problems considered and strongly affects Monte Carlo neutronics/depletion simulations.