2001 Volume 45 Pages 505-510
In the present paper, we analyze an open-channel flow over concave bed by using a particle image velocimetry (PIV) and proper orthogonal decomposition (POD). From the experimental data of velocity obtained by PIV, flow characteristics, such as mean velocity, turbulent intensity and Reynolds stress, are discussed with respect to the different Reynolds numbers. As Reynolds number becomes larger, the recirculating flow in the concave grows to a large spatial scale. The distributions of the Reynolds stress and turbulent intensity are closely related to the scale of the recirculating flow. Then, principal components of velocity fluctuations are analyzed by using POD. From the distributions of the eigenvectors, first several modes of the component relates to the effect of the recirculating flow and a few modes in succession indicate that of the coherent structures along the mixing layer. The result shows that POD is effectively available for quantitatively detecting the dominant factors of the velocity fluctuation.
