| The flow physics of a broken-down vortex impinging upon a tail is investigated by employing a technique of high-image-density particle image velocimetry. Two-dimensional velocity data are determined in different fields of view, which lead to instantaneous and averaged patterns of the flow structure. The evolution of the vortex system is interpreted with the aid of these patterns, which include distributions of velocity, vorticity, Reynolds stress and streamline topology.; During the initial stage of development of the vortex system, represented on crossflow planes, two major clusters of vortices can be observed: a primary vortex, which is the leading-edge broken-down vortex incident upon the tail; and a secondary vortex, which is generated at the tip of the tail. The circulation of the secondary vortex can be as high as approximately one-fifth the circulation of the primary vortex. The highest peaks of root-mean-square velocity and Reynolds stress occur at or very near the interface between the primary and the secondary vortices. At downstream locations along the tail, successively larger portions of the primary vortex are swept over edge of the tail. As it is severely distorted, the regions of large velocity gradient are associated with additional peaks of the root-mean-square velocity fluctuations and the Reynolds stresses.; Planes corresponding to side views of the flow structure clearly show that the highest peaks of normalized velocity fluctuation and Reynolds stress occur in the shear region between the positive and negative vortices. In addition, on planes away from the surface of the tail, additional peaks can be identified in the shear region between clusters of larger-scale vorticity.; The proper orthogonal decomposition (POD) method is applied to the PIV data to extract the most energetic flow features. The first few eigenmodes capture the major features of the primary-secondary vortex system. Smaller scale structures of the actual flow can be reconstructed with additional eigenfunctions corresponding to the instantaneous patterns of velocity and vorticity. On the other hand, it is shown that high fidelity reconstruction of the time-averaged streamline topology requires only a small number of eigenfunctions. |
