Inviscid drag prediction for transonic transport wings using a full-potential method

A theory of drag analysis in full-potential flow, based on generalization and extension of Garabedian's and McFadden's idea of determining wave drag by volume integration of the artificial viscosity, is summarized. The applicability of the theory is restricted to shock-capturi ng numerical methods (such as, e.g., finite-volume methods), where artificial viscosity is essential for proper performance. Two mesh refinement experiments on nested grids have been carried out for the DFVLR-F4-wing in transonic flow, using CH- as well as CO-topology grids. The MATRICS code used in the experiments is first-order accurate in the mesh size throughout supersonic flow regions. It is concluded that CO-topology grids are better suited for drag analysis than are CH-topology grids. It is also concluded that the accuracy of each individual drag component can be improved by extrapolating to the limit of vanishing mesh size. Finally, to avoid excessively fine grids in an engineering environment, the need is stressed for artificial viscosity terms that are second-order small in the mesh size in supersonic flow regions, except for the immediate vicinity of the shock waves. However, extrapolations procedures are believed to remain necessary for accurate drag prediction.