Assessments of the effects of several numerical parameters on RANS-based drag prediction accuracy are presented. The parameters include grid cell size adjacent to solid walls, grid stretch ratio, grid stretch transition, artificial dissipation scheme, and artificial dissipation coefficient. Results from extensive parametric studies on a two-dimensional flat plate are reported. Based on the results of these studies, guidelines for high-accuracy drag predictions using both second- and fourth-order accurate, finite-difference-based solvers are proposed. In addition, error assessments obtained with a single grid using second- and fourth-order accurate solutions are compared to multiple-grid Richardson’s extrapolation approaches. The single-grid approach is shown to provide a significant improvement in both accuracy and error assessment relative to the multiple-grid approach.
Artificial dissipation and central difference schemes for the Euler and Navier-Stokes equations
