Revealing unsteady flow structure from flow visualization images

Flow visualization images of unsteady high Reynolds number flows contain a wealth of information about the flows from which they are obtained. Unfortunately, the primary structure in the flow is often obscured by smaller structures. In addition, periodic and quasi-periodic flows often require large data sets to fully characterize them. In this article, the application of the Proper Orthogonal Decomposition (POD) to flow visualization images of a turbulent swirling jet and in the near wake behind a bluff body is considered. The results show the ability of POD to isolate the larger, more important structure. In the quasi-periodic wake flow, the results of applying POD may be used further to recreate the typical dynamic behavior of the flow. Furthermore, the modal description of the flow obtained through POD analysis provides unique insight into the unsteady flow structure and its relative importance to the flow. As a result, POD enables the extraction of information from flow visualization imagery well beyond that possible from simple inspection.