On the initialization of Rayleigh–Taylor simulations

Abstract

We demonstrate that, when initialized with experimental data, numerical simulations of Rayleigh–Taylor mixing show reasonable agreement with experimental measurements of the self-similar growth rate α. The experimental data include high-resolution velocity and density measurements from a closed water channel facility. For the simulations, a monotone integrated large eddy simulations technique was used that employed a finite-volume, Eulerian equation solver with van Leer flux limiters. Calculations were initialized with both density and velocity fluctuations. A comparison of the late-time, self-similar growth constant with the experiments showed simulations initialized with velocity perturbations gave better agreement with experiments.