Improved Thin Tube Methods for Slender Vortex Simulations

Abstract

This paper explores three numerical schemes for efficient simulation of slender vortex filaments. The schemes defeat the spatial and temporal stiffness of the equations of motion by requiring only adequate resolution of the filament centerline and allowing large integration time steps. In order to correctly capture the self-induced filament velocity, the first scheme uses an explicit velocity correction method, the second scheme relies on a logarithmic extrapolation of two velocity predictions, and the third scheme employs a local refinement algorithm. The performances of the three schemes are contrasted in light of unsteady computations of a perturbed vortex ring with small core to radius ratio.