An adaptive node refinement for particle transport techniques based on response matrix and collision probability methods

In this paper, we investigate and present an adaptive node refinement algorithm driven by an error estimator for particle transport techniques based on response matrix and collision probability methods. We represent the error estimator as a function of interface currents of the node. This indicator is used to drive a node refinement strategy for the spatial discretization. All nodes are derived from a common coarse node by hierarchic refinement. We assembled and coupled nodes from different sized nodes. Two problems featuring strong heterogeneity and highly transport streaming regime with strong flux gradients are used to test the adaptive refinement algorithm. The results show that the algorithm indeed bounds the flux error and achieves higher accuracy compared to uniform refinement.