Vortices are a critical object in turbulence research, with the primary task being the accurate identification and localization of vortex structures in complex flow fields. Vortex identification methods have evolved through three generations: the first based on vorticity and the second and third based on velocity gradient tensors. This study compares and evaluates the performance of second-generation vortex identification methods (Q, λ2, Δ, and λci) and third-generation methods (Ω, Liutex, and Ω̃R) for spanwise vortices in open-channel flow. Comparison of vortex identification rates reveals that third-generation methods exhibit significantly higher detection rates than second-generation methods. In particular, third-generation methods achieve average detection rates exceeding 78%, compared to only 47% for second-generation methods. Among third-generation methods, Ω and Ω̃R exhibit identification rates that are comparable with and higher than Liutex. Among second-generation methods, the Δ method yields the lowest identification rate (only 10%), while λci achieves the highest (up to 72%), comparable with Liutex. Analysis of vortex indicator quantities for standard and shear-contaminated vortices reveals that second-generation methods exhibit significantly elevated indicator values under shear contamination, with the Δ method yielding values hundreds of times higher than in standard vortices. In contrast, third-generation methods maintain stable or reduced indicator values, with Ω and Ω̃R showing ∼30% reductions in shear-contaminated vortices, thereby demonstrating robust resistance to shear contamination. By introducing streamwise shear and comparing changes in vortex indicator quantities, we further verify that second-generation methods are prone to elevated indicator values due to shear influence, whereas third-generation methods prove more robust. This study quantitatively confirms the superior reliability of third-generation methods in vortex identification.
He et al. (Thu,) studied this question.