Computation vs. Experiment: The Rayleigh-Taylor Instability
Maxwell Hutchinson, University of Chicago
The Rayleigh-Taylor instability drives mixing in a variety of natural and artificial systems, including oceans, nuclear reactors, and stars. Without boundaries, Rayleigh-Taylor fronts exhibit self-similar quadratic growth at late-times. Experimental and computational studies of the growth rate consistently differ by a factor of two across a wide variety of physical systems and numerical techniques. Views on the root cause of this difference mostly fall into two camps: The experiments are not reaching late-time self-similar growth, or numerical effects at small length scales over-mix the flows. We apply the Nek5000 code, which has exactly zero diffusive error, to an experimentally accessible system to address the validity of the over-mixing explanation.
Abstract Author(s): Max Hutchinson