Speaker
Description
In nuclear engineering, a criticality event is an instance in which a chain reaction reaches a state of self-sustenance. Criticality safety involves preventing inadvertent criticality events by managing factors such as material composition, geometric configuration, and neutron interactions. Thus, the quality and success of criticality safety analyses inherently depend on how well nuclear data, material distribution, and geometric parameters are known. Uncertainty propagation is therefore fundamental to the field of criticality safety. In this work, we demonstrate the Total Monte Carlo approach to geometric uncertainty propagation (the GTMC method) using mesh-based geometries. The method is implemented using the RINGEN (Random INput file GENerator) tool, which was previously used to propagate geometrical uncertainties in location and size within constructive solid geometry (CSG). The current work extends the capabilities of the RINGEN tool to incorporate mesh-based geometries, allowing it to handle other forms of geometric uncertainties such as localized deformations, shape variations, and surface roughness. We evaluate this new approach across three benchmark cases, comparing results against both adjoint-based perturbations and the original CSG-based RINGEN implementation. The GTMC method with mesh-based geometries reproduces equivalent results when modeling equivalent uncertainties, while offering vastly improved scope and fidelity for modeling intricate geometric uncertainties beyond the reach of traditional CSG. Finally, the future outlook of the GTMC method and the RINGEN tool is discussed.