Statistical Moments in Variable Density Incompressible Mixing Flows
Research Foundation of State University of New York Stony Brook
Pagination or Media Count:
Mix is critical to the modeling of chemical or nuclear reaction processes in fluids. We simulate Rayleigh-Taylor unstable pre-turbulent and transitionally turbulent fluid mixing regimes. We model experiments generated by the flow of hot and cold water over a splitter plate into an observation channel. Three statistical second moments of the flow were measured. Our simulations achieve excellent agreement with two of these and partial agreement with the third. We draw two broader lessons from this study. The first is that numerical algorithms do matter. We compare our simulations to one obtained using Miranda, a 10th-order compact stencil turbulent code. However, this code lacks front tracking, an important aspect of our simulation algorithm. The Miranda simulation misses data error bars for the measure of mix over most of the experimental time, although it is nearly DNS and uses mesh two, four and eight times finer than what we report here.
- Numerical Mathematics