Comparison of different secondary flow correction methods for depth-average flow calculations in bends

Abstract:

Several secondary flow correction methods had been proposed in literature to account for the 3d flow characteristics of secondary circulation which was often lost in depth-averaged models. In order to evaluate their performances in singular bend, successive reverse curvature bends and natural meander rivers, three different typical secondary flow correction methods were applied to a depth-averaged hydrodynamic model, which were Lien(L), Kalkwijk and Booij(K) and modified Finnie(F) methods. There are obviously limits to the descriptive ability of the three methods in simulating sharply bend channels, especially L method which showed qualitative error. As for successive reverse curvature bends and natural meander rivers, K method gave a better performance than the others. F method increased depth-averaged velocities on the outside of curves, while L method overestimated the whole flow field. By the order of magnitude analysis, we found that L method has done the maximum correction for the secondary flow and the minimum correction for transversal velocities was F method. Those three methods were evaluated by multiple performance criteria. The results of this study can provide some references for some practical applications.