Error Norm Guided Flow Analysis of Shock-Wave/Boundary-Layer Interactions

Conventional means of assessing the accuracy of Navier-Stokes analyses are difficult to implement and may not be reliable. A new approach uses direct measures of solution error, both as accuracy monitors and as guides to grid adjustment. The artificial diffusion ratio (ADR) is a promising error monitor. As an example, ADR was used to guide the Navier-Stokes analysis of a supersonic external compression inlet with bleed flow. ADR was useful for guiding grid selection and seemed to provide a measure of solution accuracy as well. The algorithm dependence of ADR and the relationship between ADR and solution accuracy were explored in a follow-on study. This study investigated the algorithm and accuracy questions in two steps. First, modified equations were developed for the MacCormack explicit and Beam-Warming implicit algorithms for Burgers equation. Second, finite difference solutions were obtained (for a range of grid densities) and solution accuracy was established through comparisons with analytic solutions. This study established that ADR was useful as guide for grid adjustment, that only a qualitative relationship exists between ADR and solution accuracy, and that this relationship is algorithm dependent. The use of ADR error monitors for general purpose applications requires additional development. Further studies are recommended to improve the understanding of the relationship between solution accuracy and artificial dissipation. Work is also recommended to develop the technique through additional model equation studies and analysis comparisons with benchmark experiments. Original contains color plates: All DTIC and NTIS reproductions will be in black and white.