Auto-correlation measurements in three-dimensional fluid flow datasets

A three-dimensional auto-correlation function is proposed as a tool to analyse 3-D fluid flow datasets. In order to test this concept, we analysed time and space evolving datasets obtained from visualizations of the flow over a half-span delta wing. The processed slices show that the regions with strong vortical motion can easily be detected. However, it is much more difficult to determine the cross-stream velocity components. Obviously, some aspects need more attention before we can decide on the usefulness of this technique: (1) the effect of resolution and noise; (2) the smoothness of the smoke field must be controlled in order to get a large grey scale value variance; (3) the time interval between two recordings must correspond to the time scale of the cross stream motions; (4) the test flow must be less complex.