Experimentally consistent large-eddy simulation of re-shocked Richtmyer–Meshkov turbulent mixing

Abstract

Re-shocked Richtmyer–Meshkov (RM) mixing is a fundamental physical process tightly related to practical mixing problems, as it involves all three classical instabilities, i.e., Rayleigh–Taylor, Richtmyer–Meshkov (RM), and Kelvin–Helmholtz instabilities. An accurate prediction of its mixing width (MW) is of significant importance. However, satisfactory prediction has not yet been achieved with the pure (not constrained by turbulence models) large-eddy simulation (LES), by which both the predicted MW and its growth rate are over-predicted. In the literature, we solve this problem by two key improvements. First, velocity perturbation, instead of the commonly used interface perturbation, is adapted to produce an initial magnitude of perturbation comparable to the corresponding experiments. Second, a new LES model is developed, with a sub-grid kinetic energy equation introduced and model coefficients dynamically determined. The key improvement here is to consider the buoyancy production effect in the closure form, which is proved to be one of the dominant mechanisms generating turbulence for the re-shocked RM mixing and remains important even at the smallest scales. Consequently, a consistent prediction of MW with experiments is realized using the present pure LES for the first time.