Incremental volumetric and Dual Kriging remapping methods

Abstract

The transfer of variables between distinct spatial domains is a problem shared by many research fields. Among other applications, it may be required for visualization purposes or for intermediate analysis of a process. In any case, two important factors must be considered: accuracy and computational performance. The accuracy becomes more important when the results have an impact on the subsequent stages of the process' analysis, as it could lead to incorrect results. The computational performance is a permanent requirement due to the ever-increasing complexity of the analysed processes. The aim of this work is to present a new remapping method, based on Dual Kriging interpolation, developed to enable accurate and efficient variable transfer operations between two different domains, discretized with hexahedral finite elements. Two strategies are proposed, which take into account different selections of interpolation points and are based on specific Finite Element Method features. They are compared with the Incremental Volumetric Remapping method in two remapping examples, one of which includes a trimming operation, highlighting their advantages and limitations. The results show that the Dual Kriging remapping method, combined with a 2D selection strategy for the donor points, can contribute to increase the accuracy of the state variables remapping operation, particularly when they present a strong gradient along the stacking direction.