Convergence improvement method for computational fluid dynamics using building-cube method

Abstract

Computational Fluid Dynamics (CFD) has become an important tool for aerodynamics by the improvements of computer performance and CFD algorithm itself. However, the computational time of CFD continues to increase, while progress of computer has been made. One of the reasons is considered that application of CFD has become more complex. For example, CFD is employed to estimate aerodynamics performance for a complex shaped object such as formula one car (Fig. 1) Concerning complex shape, however, the problem of grid generation still remains. It requires so much time and labor. To overcome the problem in meshing for complex-shaped object, we have already proposed an algorithm [3]. The algorithm consists of two approaches. One is Immersed Boundary method [6], and the other is Building-Cube Method (BCM) [4]. The basic idea of Immersed Boundary method is applied to cells in the vicinity of solid boundary, and Cartesian grid method is performed for other cells. These approaches have several advantages except for solution convergence. In this paper, Implicit Residual Smoothing (IRS) [2] is proposed for improvements of solution convergence. © 2010 Springer.