Assessment of Diffusion Operators in a Novel Moving Particle Method

Abstract

In this study, solution effectiveness is assessed to gether with its accuracy and required computational cost for three diffusion operators in a newly developed moving-particle method (MPPM) to simulate incompressible flow. The salient feature in the present particle method is the insertion of a pressure mesh within the computational particle cloud to deal with the pressure-related operators in governing equations. Three formulations to realize the velocity diffusion operator are built on particle smoothing (PS), local mesh (LM), and smoothing difference (SD) procedures. Taylor-series analyses are performed to show their formal accuracy, which is also verified by a given two-dimensional solution profile with randomly scattered particle distributions. Solution accuracy and computational cost are evaluated by solving some classic benchmark problems in comparisons with available analytical, experimental, or numerical results. Based on the evidence drawn from numerical experiment, it is concluded that the local mesh (LM) formulation is most competitive, since it can provide as accurate result as SD; while SD requires much heavier computational load to yield an accurate solution.

Acknowledgments

Financial support from the National Science Council under Grant NSC100-2221-E-022-005 is gratefully acknowledged.