Enskog kinetic theory for a model of a confined quasi-two-dimensional granular fluid

This paper derives explicit expressions for the Navier-Stokes transport coefficients of a confined quasi-two-dimensional granular gas of inelastic hard spheres at moderate densities by extending the Enskog kinetic theory via the Chapman-Enskog method and Sonine polynomial expansion.

The Gist

The Big Picture: A Dance Floor That Never Stops

Imagine a crowded dance floor where the dancers are hard, round balls (like billiard balls). In a normal fluid (like water), these dancers bounce off each other perfectly, and if you stop pushing them, they eventually stop moving.

But in granular fluids (like sand, grains of rice, or marbles), the dancers are "inelastic." Every time they bump into each other, they lose a little bit of energy. They get tired. If you don't keep pushing them, the whole dance floor freezes.

To keep the party going, someone has to inject energy. Usually, scientists simulate this by shaking the walls of the room or blowing air through the floor. But this creates a problem: the energy isn't spread out evenly. The dancers near the walls go crazy, while those in the middle are sluggish. This creates "traffic jams" and uneven flows that are very hard to predict.

The Solution in this Paper:
The authors propose a clever trick to simulate a "perfectly mixed" party. Imagine the dance floor is a very shallow box (only two balls high). The floor vibrates up and down.

• When a ball hits the vibrating ceiling or floor, it gets a little extra kick in the vertical direction.
• Because the box is so shallow, the balls can't stack up; they are forced to move sideways.
• Through collisions, that vertical "kick" gets transferred into horizontal movement.

The result? The whole dance floor stays uniformly energetic, like a perfectly mixed soup, without needing to shake the walls unevenly. The authors call this the "Delta-model."

What Did They Actually Do?

The paper is a mathematical recipe book for predicting how this "granular soup" flows. Specifically, they wanted to calculate Transport Coefficients.

Think of these coefficients as the "personality traits" of the fluid:

1. Viscosity (Sticky-ness): How hard is it to stir the soup? If it's very viscous, it's like honey; if it's low, it's like water.
2. Thermal Conductivity (Heat Sharing): How fast does a hot spot spread out? If you heat one corner, how quickly does the whole room warm up?
3. Diffusive Heat (Density Sharing): If the dancers are crowded in one spot, does that crowd affect the temperature?

The "Secret Sauce": The Delta-Model

In previous studies, scientists looked at this system only when the dancers were very far apart (a "dilute" gas). But in real life, dance floors get crowded.

This paper takes the math to the next level: Moderate Density. They asked, "What happens when the dance floor is packed, but not completely jammed?"

They used a method called Chapman-Enskog, which is essentially a way of saying: "Let's assume the dancers are mostly moving randomly, but let's add tiny corrections for when they bump into each other or when the crowd gets denser."

The Surprising Findings

When they crunched the numbers for this crowded, vibrating dance floor, they found some interesting things:

• Viscosity is Surprisingly Calm: You might think that as the dance floor gets more crowded, the fluid would get much "thicker" (more viscous). But in this specific vibrating model, the viscosity doesn't change much as you add more dancers. It's surprisingly stable.
• Heat Conductivity is Chaotic: Unlike viscosity, the ability to share heat changes wildly depending on how "bouncy" the dancers are (how much energy they lose when they hit). It goes up and down in a non-linear way.
• The "Density Heat" Effect is Tiny: There is a weird effect where density differences create heat flow. The authors found this effect is so small in this model that, for practical purposes, you can ignore it. The heat just flows from hot to cold, following the standard rules (Fourier's Law), just like in normal air or water.

Why Does This Matter?

1. Better Predictions: Engineers who design machines that handle sand, grain, or pharmaceutical powders need to know how these materials flow. This paper gives them a better math tool for when the material is moderately packed, not just when it's loose.
2. Validating the Model: It confirms that this "Delta-model" (the vibrating box trick) is a reliable way to study granular matter without the messy complications of real-world shaking.
3. Bridging the Gap: It connects the simple physics of a few marbles to the complex physics of a truckload of sand, showing us that even in a crowded, energy-losing system, there are still predictable patterns.

The Bottom Line

The authors took a complex physics problem—how to describe the flow of a crowded, energy-losing fluid—and solved it using a clever mathematical shortcut. They found that while the "stickiness" of the fluid stays surprisingly steady as it gets crowded, the way it shares heat is much more complicated. This helps scientists and engineers better understand and predict the behavior of everything from sand dunes to industrial grain silos.

Technical Summary

1. Problem Statement

Granular materials, when excited externally (e.g., by vibration), behave like fluids. However, unlike ordinary molecular fluids, grain collisions are inelastic, leading to energy dissipation. To sustain a steady state, energy must be continuously injected.

• The Challenge: Traditional methods of energy injection (e.g., vibrating walls or shear) often generate strong spatial gradients, pushing the system beyond the Navier-Stokes (hydrodynamic) regime where fluxes are linear functions of gradients.
• The Model: The authors focus on a specific theoretical model (the Delta-model) proposed by Brito et al. [25] for quasi-two-dimensional (quasi-2D) granular gases. In this setup, a granular gas is confined in a box with a vertical height smaller than two particle diameters. Vertical vibration injects energy into the vertical degrees of freedom, which is then transferred to horizontal motion via collisions.
• The Gap: Previous kinetic theory analyses of this Delta-model were limited to the low-density (Boltzmann) regime. There was a lack of theoretical description for moderate to high densities where excluded volume effects and spatial correlations become significant. The goal is to extend the theory to these densities using the Enskog kinetic equation.

2. Methodology

The authors employ the Chapman-Enskog method applied to the Revised Enskog Theory (RET) to derive hydrodynamic equations and transport coefficients.

• Collision Model:
  • Particles are smooth inelastic hard spheres with a normal restitution coefficient $\alpha \leq 1$.
  • A unique feature is the addition of an extra velocity $\Delta$ to the relative motion of colliding particles in the normal direction. This models the energy transfer from vertical to horizontal degrees of freedom.
  • The post-collisional relative velocity is modified such that the normal component is increased by a factor related to $\Delta$.
• Kinetic Equation:
  • The system is described by the Enskog kinetic equation, which includes the pair distribution function at contact, $\chi$, to account for density correlations and excluded volume effects.
  • The collision operator is adapted to include the $\Delta$ term.
• Expansion:
  • The one-particle velocity distribution function $f(\mathbf{r}, \mathbf{v}, t)$ is expanded in powers of a small parameter $\epsilon$ (related to the Knudsen number/spatial gradients).
  • Zeroth Order: Yields the local homogeneous state. The authors assume a scaled distribution function $\phi(c, \Delta^*)$ and approximate it using the leading term of a Sonine polynomial expansion (Gaussian approximation), neglecting higher-order non-Gaussian corrections (kurtosis) for simplicity, as they are found to be small.
  • First Order: Derives the constitutive equations for the pressure tensor and heat flux, identifying the Navier-Stokes transport coefficients.
• Stationary State: The analysis is particularized to the stationary temperature state, where the cooling rate due to inelasticity is exactly balanced by the energy injection from the $\Delta$ term. This allows the transport coefficients to be expressed explicitly in terms of the restitution coefficient $\alpha$ and the solid volume fraction $\phi$.

3. Key Contributions

• Extension to Moderate Densities: The paper successfully extends the kinetic theory of the Delta-model from the dilute limit to moderate densities, providing the first Enskog-based derivation of transport coefficients for this specific energy-injection model.
• Derivation of Transport Coefficients: The authors derive explicit expressions for the Navier-Stokes transport coefficients:
  • Shear viscosity ($\eta$)
  • Bulk viscosity ($\gamma$)
  • Thermal conductivity ($\kappa$)
  • Diffusive heat conductivity ($\mu$)
• Analytical Solutions: The transport coefficients are expressed as solutions to coupled linear integral equations, which are solved analytically using the Sonine approximation. The final results are presented in closed form (Table I) as functions of $\alpha$, $\phi$, and the dimensionless energy injection parameter $\Delta^*_s$.
• Validation of Approximations: The authors demonstrate that the non-Gaussian corrections (kurtosis $a_2$) to the velocity distribution are small ($|a_2| \lesssim 0.1$) in the stationary state, justifying the use of the Gaussian (Maxwellian) approximation for the zeroth-order distribution in calculating the transport coefficients.

4. Key Results

The paper presents the scaled transport coefficients ($\eta^*, \kappa^*, \mu^*, \gamma^*$) for a 2D system. Key findings include:

• Shear Viscosity ($\eta^*$):
  • Surprisingly, the scaled shear viscosity exhibits a weak dependence on density ($\phi$) compared to other granular models (like the standard Inelastic Hard Sphere model).
  • While density corrections exist, they largely cancel out or are small, meaning the viscosity is primarily determined by the restitution coefficient $\alpha$.
  • This result aligns well with Molecular Dynamics (MD) simulations for both dilute and dense systems.
• Thermal Conductivity ($\kappa^*$):
  • The scaled thermal conductivity shows a non-monotonic dependence on the restitution coefficient $\alpha$.
  • Unlike viscosity, the thermal conductivity is more sensitive to density and dissipation.
• Diffusive Heat Conductivity ($\mu^*$):
  • This coefficient, which relates heat flux to density gradients, is generally negative and small in magnitude.
  • Crucially, for practical purposes, the contribution of the density gradient to the heat flux is negligible. Thus, the heat flux obeys Fourier's Law ($\mathbf{q} = -\kappa \nabla T$) even in this driven, inelastic system, contrasting with undriven granular gases where $\mu$ can be significant.
• Stationary Temperature: The model predicts a stable stationary temperature where energy injection balances dissipation, dependent on $\alpha$ and $\Delta$ but largely independent of density.

5. Significance

• Theoretical Foundation: This work provides the necessary theoretical framework to apply the Delta-model to real-world quasi-2D granular experiments (e.g., vibrated monolayers) at moderate densities, bridging the gap between low-density theory and experimental conditions.
• Predictive Power: The explicit formulas for transport coefficients allow for direct quantitative comparison with Molecular Dynamics simulations and experimental data, validating the kinetic theory approach for driven granular fluids.
• Hydrodynamic Stability: The derived coefficients are essential for analyzing the linear hydrodynamic stability of the homogeneous steady state. The authors note that the positive definiteness of the hydrodynamic matrix suggests the steady state remains stable in the dense regime, unlike some other driven models that exhibit phase separation.
• Practical Application: By confirming that Fourier's law holds and providing accurate viscosity/conductivity values, the paper enables more reliable modeling of granular flows in industrial and geophysical contexts where quasi-2D confinement and vibration are present.

In summary, this paper rigorously extends the kinetic theory of driven granular gases to moderate densities, revealing that while density effects are present, the shear viscosity remains surprisingly robust, and the heat flux behavior simplifies to standard Fourier conduction under these specific driving conditions.