Filament surgery and temporal grid adaptivity extensions to a parallel tree code for simulation and diagnosis in 3D vortex dynamics

Abstract

The vortex collapse-reconnection process presents behaviors commonly observed in turbulent flows: multiple spatial and temporal scales, rapid vorticity and strain-rate amplification and dissipation through generation of small scales. In this work we reduce the computational complexity of our problem by using hierarchical methods (tree codes), introducing a time extrapolation framework for each particle, and applying a filament algorithm, based on the energy density, as a regularization for vortex collapse. The high performance parallel implementation of Barnes-Hut algorithm permit us to increase by one order of magnitude the resolution of the vortex collapse simulations. The use of the time extrapolation for slow moving particles helps in concentrating the computational effort in the important dynamic domains. The vortex filament surgery regularizes effectively the growth of the number of particles in the collapse regions of the flow. The reduction in complexity achieved will contribute to optimize the use of the numerical simulations in the reduced model building process.