Data mining of coherent structures from turbulence tomographic TR-PIV signal database utilizing improved quadrant splitting method

In this paper, we calculated the spatial locally averaged velocity strains along the streamwise direction with four scales in the velocity vector field of turbulent boundary layer by utilizing the three-dimensional three-component database of time series of velocity field measured by tomographic time-resolved particle image velocimetry. We introduced an improved quadrant splitting method based on the spatial local-averaged velocity strains and a new conditional sampling phase average method as a data-mining criterion to detect and extract the inherent coherent structures from such turbulence signal database. Furthermore, we used the improved quadrant splitting method to obtain and investigate the spatial topologies of fluctuating velocity and fluctuating vorticity with four scales whose center was a strong second-quadrant event (Q2) or a fourth-quadrant event (Q4). Results illustrate that there is a better similarity of the multi-scale coherent structures in wall-normal direction than the one in other two dimensions.