Evaluation of Some Data Transfer Algorithms for Noncontiguous Meshes

Abstract

The objective of this research was to identify and evaluate selected suitable methods to transfer information between computational methods with noncontiguous meshes or grids. This data transfer can easily be the limiting factor in the accuracy of computational simulations in a variety of applications. The data to be transferred can include point variables, such as deflections, pressure, and temperature; area-based variables, such as loads; and rates, such as heat flux. A method should provide smooth, yet accurate, transfer of data for a wide variety of functional forms. An extensive literature survey was completed that identified current algorithms in aerospace applications, as well as other candidate algorithms from different disciplines, such as mapping and CAD/CAM. The performance of the various methods was assessed by a series of test cases, including the mapping of constant and linear functions, as well as sinusoidal functions with varying numbers of oscillations within the domain. Accuracy, computational memory requirements, and computational time requirements were all evaluated.