Tracer-Particle Advection: Algorithm Components and Implementation Methods

Tracer-particles, massless particles that are advected throughout the flow domain, are an important factor in computational solutions. The utility of tracer-particles, which sample and report relevant solution data, is predicated upon an accurate advection algorithm. The objective of this research was to investigate the ability of existing techniques, physical-space and logical-space advection, to accurately predict tracer-particle pathlines within various grid topologies. The physical-space technique, using a higher-order integration method, accurately predicted pathlines for a curved flow-field within a nonorthogonal grid. In contrast, the logical-space technique failed to accurately predict pathlines for a uniform flow-field within a nonuniform rectilinear grid. Existing logical-space advection techniques are, therefore, limited to uniform rectilinear grids. Introduction A primary goal of computational simulation research is the development of accurate and efficient methods to solve physical models. The full power and utility of simulation codes requires the ancillary ability to perform insightful data interpretation [1]. Data interpretation is used to examine the solution and also as a diagnostic tool to improve the computational technique [2]. The purpose of this research was to investigate tracer-particle methods which are used extensively for data extraction within material advection models. Material advection, the forced motion of solid or fluid material, is a fundamental physical process and is the core of many simulation codes. Tracer-particles, massless particles that are advected throughout the computational domain concurrent with the solution evolution, are used to sample and report relevant data. The utility of tracer-particles is predicated upon an accurate and efficient advection algorithm [3]. Some popular post-processing tools [4] generate steady-state streamlines, which obviates a tracer-particle capability within time-independent simulation codes. For time-dependent simulations, some post-processing tools [7] generate unsteady tracer-particle pathlines by interpolating velocity flow-field data between solution time-steps. However, this introduces addition numerical approximations within the advection algorithm, and it may reduce the accuracy of the tracer-particle pathline. Therefore, reliance on a post-processing tool for accurate unsteady pathline evolution is questionable. In addition, the exclusion of a pathline capability within the simulation code limits the utility of tracer-particles as a model diagnostic tool. A tracerparticle capability is then recommended for inclusion within all modern simulation codes. Two types of tracer-particle advection are recognized: physical-space and logical-space