🔢 mathematics

Derivation of effective kinetic equations describing oscillations in viscoelasticity and in compressible Navier-Stokes

This paper presents lecture notes on constructing solutions with persistent oscillations for hyperbolic-parabolic systems in viscoelasticity and compressible Navier-Stokes, demonstrating how kinetic formulations can be used to derive effective equations coupling kinetic dynamics with macroscopic flows.

Original authors: Athanasios E. Tzavaras

Published 2026-04-16

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Athanasios E. Tzavaras

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Picture: The Problem of the "Flickering" Fluid

Imagine you are watching a crowd of people in a stadium.

Scenario A (Smooth): Everyone stands still. The crowd is calm.
Scenario B (Waves): Everyone does "The Wave." It moves smoothly across the stadium.
Scenario C (The Chaos): Imagine a crowd where people are frantically jumping up and down, but not in a wave. Instead, they are jumping in a chaotic, high-speed pattern that never settles down. If you stand far away and squint, the crowd looks like a fuzzy, vibrating blur. You can't see individual people; you just see a "fuzz" of motion.

This paper is about Scenario C.

In physics, we often study fluids (like air or water) and materials (like rubber or metal). Usually, we use equations to predict exactly how they move. But sometimes, especially when materials change phases (like ice melting into water) or when gases are under weird pressure, the math gets messy. The solution doesn't settle into a smooth pattern. Instead, it develops persistent oscillations—tiny, rapid, chaotic fluctuations that never die out, even as time goes on.

The author asks: "If the solution is just a fuzzy blur of chaos, how do we write a new set of rules to describe that blur?"

The Two Main Characters

The paper looks at two specific types of "fuzzy" systems:

Viscoelasticity (The Stretchy Rubber): Think of a piece of chewing gum or a memory foam pillow. When you pull it, it stretches. Sometimes, depending on the material, it wants to snap back to two different shapes at once (like a double-well potential). If you pull it, it might start vibrating wildly between these two shapes.
Compressible Navier-Stokes (The Squeezable Gas): Think of a gas in a piston. If you squeeze it, the pressure goes up. But if the gas behaves strangely (like a Van der Waals gas), squeezing it might make the pressure drop, then rise, then drop again. This instability causes the density of the gas to jump back and forth rapidly, creating a "fuzzy" density field.

The Old Way vs. The New Way

The Old Way (Young Measures):
Mathematicians have a tool called "Young Measures." Imagine you take a photo of that chaotic crowd and count how many people are jumping at height $X$ , height $Y$ , etc. You get a probability distribution.

The Problem: This tells you what the crowd looks like statistically, but it doesn't tell you how that crowd moves or evolves over time. It's like having a photo of the blur, but no video of the action. It's a static snapshot of chaos.

The New Way (Kinetic Equations):
The author proposes a new tool: Kinetic Equations.
Instead of just counting the people, imagine we give every single person in the crowd a tiny GPS tracker that records their specific height and speed. We then write a rule that describes how the entire distribution of these trackers moves.

Think of it like this:

Old View: "The crowd is 50% jumping up and 50% standing still."
New View (Kinetic): "The 'jumping-up' part of the crowd is moving to the left, while the 'standing-still' part is moving to the right, and they are swapping places based on the stress of the material."

The "Secret Sauce": The Distribution Function

The paper introduces a specific mathematical object called a Kinetic Function (let's call it $F$ ).

Imagine $F$ is a sliding window or a volume knob for the chaos.

If you look at a specific point in space and time, $F$ tells you: "What percentage of the material at this spot has a strain (stretch) less than value $\xi$ ?"
As time passes, this "window" slides and changes shape.

The author derives a new equation (Equation 1.8 in the paper) that describes how this window moves.

It looks like a traffic flow equation.
The "cars" in this traffic are the different possible states of the material (e.g., "stretched a little," "stretched a lot").
The "traffic lights" are the physical forces (stress, pressure).
The equation tells us how the "traffic" of these different states rearranges itself over time.

The "Double-Well" Analogy

To understand why this is hard, imagine a ball in a landscape with two valleys (a double-well potential).

Valley A: A relaxed state.
Valley B: A stretched state.
The Hill: An unstable peak in the middle.

If you shake the ball (add energy), it might bounce back and forth between Valley A and Valley B very fast.

Standard Math: Tries to average this out. "Okay, the ball is 50% in A and 50% in B." But this average is a point on the hill, which is physically impossible! The ball is never actually on the hill.
This Paper's Math: Says, "Don't average it. Track the probability of the ball being in A vs. B." It creates a map showing how the "probability cloud" moves. If you push the system, the cloud might split, with some part staying in A and some part rushing to B.

Why Does This Matter?

Phase Transitions: This helps us understand how materials change states (like water freezing or metal changing crystal structures) without the math breaking down.
Turbulence: It offers a new way to look at turbulence in fluids, where things get chaotic and "fuzzy."
Better Simulations: Engineers can use these new "Effective Kinetic Equations" to simulate materials that are too chaotic for standard computers to handle. Instead of trying to calculate every tiny vibration, they calculate the movement of the "fuzz" itself.

Summary in One Sentence

This paper teaches us how to stop trying to smooth out the chaotic, vibrating mess of complex materials and gases, and instead write a new set of "traffic rules" that describe exactly how that mess moves and evolves over time.

1. Problem Statement

The paper addresses the mathematical challenge of nonlinear homogenization for hyperbolic-parabolic systems that exhibit persistent oscillations. These oscillations typically arise in mechanical systems involving:

Phase transitions: Modeled by viscoelastic materials with non-convex strain energy potentials (double-well potentials).
Compressible fluids: Modeled by the compressible Navier-Stokes equations with non-monotone pressure laws (e.g., Van der Waals gases).

The Core Difficulty:
In standard weak convergence theory, if a sequence of solutions $(u_\epsilon, v_\epsilon)$ oscillates, the weak limit $(u, v)$ often fails to satisfy the original nonlinear equations because the limit of the nonlinear terms (e.g., $\sigma(u_\epsilon)$ or $p(\rho_\epsilon)$ ) does not equal the nonlinear function of the limit ( $\sigma(u)$ or $p(\rho)$ ).

Young Measures: Traditionally used to describe the statistical distribution of these oscillations. However, Young measures are static probability distributions that do not inherently capture the dynamics of how oscillations propagate or evolve over time.
The Goal: To derive effective kinetic equations that couple a macroscopic flow with a kinetic equation describing the evolution of the oscillation distribution, thereby providing a dynamic description of the homogenized system.

2. Methodology

The author employs a methodology rooted in the kinetic formulation of scalar conservation laws (Lions-Perthame-Tadmor) and extends it to systems. The key steps include:

Construction of Oscillatory Solutions:
- The paper first constructs explicit examples of solutions with persistent oscillations for both linear and nonlinear systems.
- Viscoelasticity: Uses non-monotone stress functions $\sigma(u)$ to create standing shock waves where the strain $u$ oscillates between two phases while velocity $v$ remains continuous.
- Navier-Stokes: Constructs solutions where density $\rho$ oscillates across characteristic surfaces while velocity $u$ remains smooth (e.g., $u = y/t$ ).
Transition from Young Measures to Kinetic Functions:
- Instead of working directly with the Young measure $\nu_{t,x}(\lambda)$ , the author utilizes its distribution function (or kinetic function) $F(t, x, \xi)$ .
- For a sequence $u_\epsilon$ , the kinetic function is defined as the weak limit of the indicator function:
  $1_{\{u_\epsilon < \xi\}} \rightharpoonup F(t, x, \xi)$
- This function $F$ acts as the cumulative distribution function of the Young measure: $F(t, x, \xi) = \nu_{t,x}((-\infty, \xi])$ .
Derivation of Effective Equations:
- By applying the chain rule to the indicator function and passing to the limit in the weak formulation of the original PDEs, the author derives a transport equation for $F$ .
- This transport equation is coupled with the macroscopic momentum/conservation equations, where the nonlinear terms are replaced by moments of the kinetic function (e.g., $\int \sigma(\xi) dF$ ).

3. Key Contributions and Results

A. Existence and Oscillation in Linear/Nonlinear Systems

Linear Systems: Demonstrates that in hyperbolic-parabolic systems with singular diffusion matrices (where viscosity acts only on some variables, e.g., velocity but not strain), high-frequency oscillations in the initial data can persist in the solution. Specifically, while velocity and temperature decay, strain and strain-rate oscillations remain.
Nonlinear Phase Transitions: Proves the existence of "twinning" solutions in Kelvin-Voigt viscoelasticity. These are dynamic solutions where the deformation gradient jumps between two states across moving interfaces, creating sustained oscillations.
Compressible Navier-Stokes: Constructs periodic solutions for barotropic fluids with non-monotone pressure where density jumps across characteristic lines while velocity remains continuous.

B. Derivation of Effective Kinetic Equations for Viscoelasticity (1D)

For the system describing shear motions:
$\partial_t u = \partial_x v, \quad \partial_t v = \partial_x (\sigma(u) + \partial_x v)$
The author derives the effective system (Theorem 7.3) governing the limit $(v, F)$ :

Kinetic Equation:
$\partial_t F + \partial_\xi \left( (v_x + \sigma(u) - \sigma(\xi)) F \right) + \sigma'(\xi)F = 0$
Macroscopic Momentum:
$\partial_t v = \partial_{xx} v + \partial_x \left( \int \sigma(\xi) dF_{t,x}(\xi) \right)$
Stress Closure:
$S = \sigma(u) + v_x = \int \sigma(\xi) dF_{t,x}(\xi) + v_x$

Significance: This system separates phases. If initial data crosses an unstable region of the stress-strain curve, the kinetic equation predicts the evolution of the distribution $F$ toward stable equilibria (phase separation), effectively describing the homogenization process dynamically.

C. Derivation of Effective Kinetic Equations for Compressible Navier-Stokes

For the system:
$\partial_t \rho + \text{div}(\rho u) = 0, \quad \partial_t (\rho u) + \text{div}(\rho u \otimes u) + \nabla p(\rho) = \text{div}(\mathbb{S})$
The author outlines two approaches to derive an effective equation for the density oscillations:

Approach 1 (Renormalized Solutions): Uses the distribution function of the Young measure for density, $F(t, x, \xi)$ . The resulting equation involves the effective viscous flux and is complex due to the coupling with the pressure term.
Approach 2 (Weighted Kinetic Function): Introduces a weighted function $H(t, x, \xi) = \text{weak-lim}(\rho_\epsilon 1_{\{\rho_\epsilon < \xi\}})$ $H (t, x, ξ) = weak-lim (ρ_{ϵ} 1_{{ρ_{ϵ} < ξ}})$ .
- The derived effective equation (Equation 8.18) is:
  $\partial_t H + \text{div}(uH) - \frac{1}{\lambda+2\mu}\xi \left( p(\xi) - p(\rho) + (\lambda+2\mu)\text{div}u \right) \partial_\xi H = 0$
- This is coupled with the standard momentum equation where the pressure term is replaced by the moment $\int p(\xi) dH$ .

4. Significance and Impact

Dynamic Homogenization: The primary contribution is moving beyond static Young measures to dynamic kinetic equations. This allows researchers to track how oscillations evolve, merge, or separate over time, rather than just knowing their statistical distribution at a snapshot.
Phase Transition Modeling: The derived kinetic equations provide a rigorous mathematical framework for modeling phase separation in solids (viscoelasticity) and fluids (non-monotone gases) where standard continuum mechanics fails due to the lack of convexity in the energy/pressure functions.
Unification of Methods: The paper successfully bridges the gap between the theory of scalar conservation laws (where kinetic formulations are well-established) and complex systems of conservation laws (viscoelasticity and Navier-Stokes).
Resolution of Oscillation Cancellation: The work provides a framework to understand when oscillations persist (due to non-convexity) versus when they cancel out (due to entropy conditions), offering a tool to analyze turbulence and multi-phase flows.

5. Conclusion

Tzavaras's lecture notes provide a comprehensive roadmap for deriving effective equations for hyperbolic-parabolic systems with persistent oscillations. By replacing the Young measure with its distribution function (kinetic function), the author transforms the homogenization problem into a coupled system of a macroscopic PDE and a kinetic transport equation. This approach offers a powerful new tool for analyzing phase transitions and compressible fluid dynamics where traditional weak convergence methods are insufficient to capture the dynamics of the oscillating microstructure.