Statistics of mixing in three-dimensional Rayleigh-Taylor turbulence at low Atwood number and Prandtl number one

Three-dimensional miscible Rayleigh-Taylor (RT) turbulence at small Atwood number and at Prandtl number one is investigated by means of high resolution direct numerical simulations of the Boussinesq equations. RT turbulence is a paradigmatic time-dependent turbulent system in which the integral scale grows in time following the evolution of the mixing region. In order to fully characterize the statistical properties of the flow, both temporal and spatial behaviors of relevant statistical indicators have been analyzed. Scaling of both global quantities (e. g., Rayleigh, Nusselt, and Reynolds numbers) and scale dependent observables built in terms of velocity and temperature fluctuations are considered. We extend the mean-field analysis for velocity and temperature fluctuations to take into account intermittency, both in time and space domains. We show that the resulting scaling exponents are compatible with that of the classical Navier-Stokes turbulence advecting a passive scalar at comparable Reynolds number. Our results support the scenario of universality of turbulence with respect to both the injection mechanism and the geometry of the flow. (C) 2010 American Institute of Physics. [doi: 10.1063/1.3371712]