Structural changes in fluid flows are commonly observed in phenomena such as vortex formation, breakdown, and separation. This study presents a data-driven framework for quantitatively assessing similarity in flow dynamics by integrating Dynamic Mode Decomposition (DMD), which extracts dominant flow features, with a Grassmann kernel-based metric. Simulations of flow over an airfoil were conducted across a range of angles of attack, and changes in the lift coefficient and DMD modes were examined. The similarity analysis revealed a sharp decline near an angle of attack of 16°, corresponding to the onset of stall associated with boundary layer separation, thereby providing a quantitative characterization of the transition in the flow structure. The proposed approach offers a practical and interpretable tool for analyzing and comparing complex fluid dynamic behaviors.
Nakasone et al. (Wed,) studied this question.