TL;DR
This paper presents a novel statistical mechanics framework to predict mixing efficiency in stratified fluids. By treating turbulent stirring as a process that drives the system towards an equilibrium state, the authors derive a method to quantify how much injected energy contributes to potential energy increase (mixing) versus dissipation. This approach offers a theoretical foundation for understanding mixing in scenarios where deterministic modeling is intractable.
What Problem Does This Paper Address?
Predicting the efficiency of mixing in stratified fluids, particularly in turbulent scenarios, is a long-standing challenge. When energy is injected into such a fluid, it can lead to mixing (increasing potential energy by rearranging density layers) or be dissipated as kinetic energy. The sheer number of degrees of freedom in turbulent flows makes deterministic, first-principles calculations incredibly difficult. This paper aims to bridge this gap by applying statistical mechanics to predict the cumulative mixing efficiency, which represents the total amount of mixing achieved over time, as a function of key parameters like the global Richardson number and the initial stratification profile.
Methods
The authors employ an equilibrium statistical mechanics approach. The core idea is that an inviscid, adiabatic Boussinesq fluid subjected to turbulent stirring will irreversibly evolve towards an equilibrium state. This equilibrium state is characterized by:
- Smooth, coarse-grained buoyancy profile: The large-scale stratification becomes smoother.
- Small-scale velocity and buoyancy fluctuations: These fluctuations carry kinetic energy and represent the "mixed" state.
The theory involves:
- Identifying phase-space variables: Velocity and buoyancy fields are identified as suitable variables that satisfy Liouville's theorem, ensuring equiprobability of microscopic configurations.
- Discretization: The continuous fields are discretized to allow for counting microscopic configurations.
- Macroscopic description: A probability distribution function (PDF) is used to describe the macroscopic state, representing the frequency of buoyancy and velocity levels within a given region.
- Variational problem: The equilibrium state is found by maximizing the macrostate entropy subject to constraints of conserved total energy and global buoyancy distribution.
- Relaxation algorithm: A numerical algorithm based on a maximum entropy production principle is developed to compute equilibrium states for arbitrary background buoyancy profiles.
The mixing efficiency ($\eta$) is defined as the ratio of the increase in available potential energy ($E_p$) to the total injected energy ($E_{inj}$), where $E_{inj} = E_p + E_c$ (with $E_c$ being the kinetic energy of fluctuations).
Datasets
Not discussed in the paper. The paper focuses on theoretical development and numerical computations based on defined initial conditions and parameters, rather than empirical datasets.
Key Insights and Ideas
- Irreversible Mixing via Stirring: The theory demonstrates that even in an inviscid and adiabatic system, turbulent stirring leads to an irreversible increase in potential energy, effectively mixing the fluid at a coarse-grained level. This is a crucial insight, as it decouples mixing from molecular diffusion in the idealized limit.
- Equilibrium as an Attractor: The equilibrium state, characterized by a smooth buoyancy profile and small-scale fluctuations, acts as an attractor for the dynamics. This allows for the prediction of the final outcome of mixing, even if the relaxation process itself is not explicitly modeled by the equilibrium theory.
- Mixing Efficiency Depends on Background Profile: A key finding is that mixing efficiency is not solely determined by the global Richardson number but also strongly depends on the shape of the background buoyancy profile. This is a significant departure from simpler models.
- Asymptotic Mixing Efficiency of 0.25: For a strictly monotonic background buoyancy profile and high Richardson numbers, the theory predicts an asymptotic mixing efficiency of 0.25, supported by energy equipartition arguments. This provides theoretical backing for previous phenomenological observations.
- Two-Level vs. Linear Stratification: The paper highlights contrasting behaviors for different background profiles: a bell-shaped mixing efficiency for a two-level stratification and a monotonic increase towards 0.25 for a linear stratification.
Key Results
- A statistical mechanics framework is developed to predict cumulative mixing efficiency ($\eta$) in stratified fluids.
- The theory predicts that $\eta$ increases linearly with the Richardson number (Ri) in the limit of weak Ri, irrespective of the background buoyancy profile.
- For a strictly monotonic background buoyancy profile and large Ri, $\eta$ approaches 0.25.
- For a two-level background buoyancy profile, $\eta$ exhibits a bell shape with a maximum around 0.15.
- The paper provides a numerical algorithm to compute equilibrium states and mixing efficiency for arbitrary background buoyancy profiles.
Limitations and Future Work
- Inviscid and Adiabatic Assumption: The theory is derived for inviscid and adiabatic Boussinesq fluids. Its applicability to real-world scenarios is limited to the regime of high Reynolds and Péclet numbers, where viscous and diffusive effects are secondary. The paper acknowledges that the system may not fully explore phase space in forced-dissipative configurations.
- Ergodicity Assumption: The statistical mechanics approach relies on the assumption that the system explores phase space evenly (ergodicity). This is favored by large-scale stirring but may not hold for turbulence driven by small-scale forcing, which can lead to local mixing before large-scale stirring.
- No Prediction of Dynamics: The theory predicts the final equilibrium state but not the temporal evolution or the dynamics of the relaxation process. It does not predict turbulent diffusivity or viscosity during the relaxation phase.
- Future Work: The authors suggest that the relaxation equations used in their algorithm could be adapted to model relaminarization after a mixing event. They also propose that this approach could offer a complementary perspective to other statistical or stochastic methods used in fields like combustion. Further investigation into the role of injection mechanisms and their impact on ergodicity is also suggested.
Summary updated