Abstract: Nous présentons des mesures quantitatives de fluctuations de densité dans un fluide bi-couche mélangé par une grille oscillante. La turbulence produite par la grille localisée dans la couche inférieure dense diffuse rapidement dans cette couche puis est bloquée à l'interface entre le fluide dense et le fluide léger. La hauteur de l'interface augmente alors lentement en entrainant du fluide de la couche supérieure dans la couche inférieure. Nous proposons une méthode qui permet de séparer une partie ondulatoire d'une partie turbulente dans les fluctuations de densité à l'interface. Ces fluctuations sont observées par la méthode de fluorescence induite par un plan laser. Nous interprétons ces résultats à l'aide d'un modèle phénoménologique basé sur une approche statistique du mélange.
Introduction
In geophysical flows such as the ocean and atmosphere, fluids are frequently stratified—organized into layers of varying density (e.g., warm water sitting on cold water). Mixing between these layers is a fundamental driver of global climate patterns, yet it occurs at scales significantly smaller than those resolved by current global simulation models.
To improve these models, researchers require accurate "sub-grid" physics describing how turbulence in one layer "entrains" (pulls in) fluid from an adjacent layer. A primary challenge in experimental fluid dynamics is distinguishing between large-scale interface oscillations (internal waves) and the small-scale chaotic fluctuations that drive actual mixing.
Experimental Methodology
This study, presented at the 21st French Congress of Mechanics (CFM 2013), utilizes a controlled laboratory environment to investigate these dynamics.
Laboratory Setup
The experiments were conducted in a 40cm x 40cm tank containing a two-layer fluid system. The bottom layer consisted of dense salt water dyed with rhodamine, while the top layer was a lighter mixture of water and ethanol. A critical technical requirement was matching the refractive indices of both layers to ensure that the laser sheet could penetrate the interface without optical distortion.
Turbulence and Imaging
Turbulence was generated by an oscillating grid located at the bottom of the tank. This turbulence propagated upward until encountering the density interface. The researchers employed Planar Laser Induced Fluorescence (PLIF) to capture high-resolution 2D density maps. By measuring the fluorescence intensity of the dye, the exact concentration of the fluids was calculated at every pixel.
Statistical Analysis and Key Insights
The study analyzed data from seven experimental runs with varying density differences ($\Delta\rho$), ranging from 0.01% to 1.3%.
The "Tanh" Universality
A significant finding of this research is that, despite the inherent chaos of turbulence, the average density profiles across all experiments collapsed onto a universal mathematical curve: the hyperbolic tangent ($tanh$). This suggests that the mixing zone reaches a predictable statistical equilibrium.
Separating Waves from Turbulence
The authors developed a robust method to isolate large-scale interface height variations ($h(x, t)$) from small-scale turbulent fluctuations. This separation is vital for accurately calculating entrainment velocities and understanding the underlying physics of stratified flows.
The Role of the Richardson Number
The Richardson number ($Ri$), representing the ratio of potential energy to kinetic energy, was identified as the primary governing parameter. The results suggest two distinct mixing regimes depending on $Ri$, which determines the efficiency with which turbulence overcomes gravity to mix the layers.
Conclusion and Future Directions
The experimental data showed excellent agreement with phenomenological models based on statistical physics. However, several areas remain for future investigation:
- Entrainment Velocity: Finalizing the mathematical relationship between entrainment speed and the Richardson number.
- Filamentation: Investigating how thin "filaments" of fluid pulled across the interface create asymmetry in concentration distributions.
- Shear-Driven Turbulence: Determining if these universal statistical laws hold for turbulence driven by shear, such as wind-driven mixing in the ocean.
This work provides a critical step toward more accurate parameterizations of mixing in global climate and weather models.
Summary updated