| We examine experimentally and describe theoretically the properties of turbulence developing from buoyancy-driven and shear-driven instabilities that can occur at the interface between two fluids. In particular, we examine those circumstances found in our experimental and numerical investigations that depart from spectral equilibrium. To calculate the stochastic behavior of turbulence, it is necessary to model the equations governing the correlations of turbulent fluctuations, which interact with pressure gradients and/or mean-flow shear. To evaluate the suitability of a model for representing the physics of a very complex flow, it is necessary to have experimental data for comparison with calculated results. Turbulence is comprised of many scales, or spectra, of correlated motion; the dominant, or large scales receive most of their energy from the mean-flow kinetic and/or potential energy. Quantifying the rate of energy transference from large scales to small-scale motion is one of the problems that all models have to address. Most models (e.g., the K-ϵ model) assume self-similarity for the turbulent spectra, and that the rates of cascade in these spectra are governed solely by the dominant scales—an assumption which can be violated in a variety of circumstances. The principal source of violation that we investigate occurs when there is a combination of shear and buoyancy-driven turbulence. |
