A Local Adaptive Grid Refinement Strategy for Block-Structured Finite-Volume Solvers

摘要

A major problem in computational fluid dynamics (CFD) is the occurence of huge scale differences when solving the underlying partial differential equations, so different grid resolution in different parts of the computational domain id needed. This leads to high memory requirements and large cpu times. As a solution for this problem we present a local adaptive grid refinement strategy that is combined wirh a multigrid method. We describe the methodology and its implementation in FASTEST3D, a finite volume flow solver. Two examples are discussed to illustrate the approach.