Performance Analysis of Parallel Algebraic Preconditioners for Solving the RANS Equations Using Fluctuation Splitting Schemes

Abstract

We consider iterative solution strategies for solving the Reynolds-Favre averaged Navier-Stokes (RANS) equations on 2D and 3D flow configurations. The novelty of this study is the coupling of an hybrid class of methods for the space discretization, called Fluctuation Splitting (or residual distribution) schemes, and a fully coupled Newton algorithm for solving the RANS equations. This approach is particularly attractive for parallel computations because it gives rise to discretization matrices with a compact stencil resulting in a limited number of nonzero entries. In this paper, we present the solution approach and report on results of numerical experiments with particular emphasis on the design of preconditioners for the inner linear system, which is a critical computational issue of the iterative solution.