Separation of Intertwined Vortical Structures in Turbulent Channel Flow Using Contour Tree-Based Segmentation

Vortices are fundamental to turbulent flow dynamics, often forming complex and intertwined configurations, especially at high Reynolds numbers. Traditional region-based vortex extraction methods (e.g., λ₂, Q, λ_ci, Rortex) are threshold-sensitive and struggle to differentiate individual vortices within intertwined vortical regions. Building on our previous work, where we developed a toolset for identifying and extracting individual vortices by exploring their spatial hierarchical representation and characterizing them based on their physical attributes (e.g., vorticity, enstrophy, velocity) and geometric information, we now employ a Contour Tree-based segmentation (CT) approach with an additional ‘layering’ step, aiming to improve the accuracy of separating vortices in complex regions. Traditional CTs utilize scalar field critical points for segmentation, which may overly separate vortices. To address this, we incorporate vorticity lines, to help determine whether two adjacent regions belong to one vortex or not. We demonstrate the effectiveness of our method by applying it to the Channel Flow DNS datasets from the Johns Hopkins Turbulence Database (JHTDB).