🔢 mathematics

On the passage from nonlinear to linearized viscoelastodynamics

This paper rigorously derives the equations of linearized viscoelastodynamics in Kelvin-Voigt rheology from a nonlinear, frame-indifferent model and proves the convergence of both continuous and time-discrete solutions between the nonlinear and linearized systems.

Original authors: Barbora Benešová, Malte Kampschulte, Martin Kružík

Published 2026-03-04

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Barbora Benešová, Malte Kampschulte, Martin Kružík

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

Imagine you are holding a piece of jelly. If you poke it gently, it wobbles and slowly settles back into shape. If you poke it hard and fast, it might bounce, vibrate, or even tear.

This paper is about figuring out how to mathematically describe that wobble. Specifically, the authors are trying to connect two very different ways of looking at the same problem:

The "Real World" View (Nonlinear): This is the messy, complicated reality where the jelly stretches, twists, and changes shape in complex ways. The math here is incredibly hard, like trying to solve a puzzle where the pieces keep changing shape while you're holding them.
The "Simplified" View (Linear): This is the "textbook" view where we assume the jelly only moves a tiny bit. The math here is much easier, like a straight line instead of a squiggly curve.

The big question the authors ask is: "Can we prove that the messy, complicated math actually turns into the simple, textbook math when the forces are small?"

Here is the breakdown of their journey, using some everyday analogies.

1. The Three Lenses (The Scales of Time)

The authors realize that how we look at the jelly depends on how fast we are watching it. They use three different "lenses" or time scales:

The Hyper-Fast Lens (Dynamic/Inertial): This is like watching the jelly in slow motion with a high-speed camera. You see every bounce, every vibration, and the "inertia" (the jelly's desire to keep moving) is huge. This is the Nonlinear Dynamic model.
The Slow-Motion Lens (Quasi-Static): This is like watching the jelly settle after you stop poking it. The bouncing has stopped, but it's still slowly sliding into its final shape. The "inertia" is ignored because things are moving so slowly. This is the Nonlinear Quasi-Static model.
The Frozen Lens (Static): This is looking at the jelly after it has completely stopped moving. It's just sitting there. This is the Static model.

The Analogy: Imagine a car driving down a hill.

Dynamic: The car is speeding, hitting bumps, and the engine is revving (inertia matters).
Quasi-Static: The car is coasting very slowly down a gentle slope; you barely feel the engine.
Static: The car is parked at the bottom.

2. The Magic Trick: "Time-Delayed" Steps

The authors use a clever trick to connect these lenses. They imagine taking the "Hyper-Fast" view and breaking time into tiny, tiny steps (like frames in a movie).

They introduce a "Time-Delay" concept. Imagine you are trying to predict where the jelly will be next second. Instead of guessing based on now, you look at where it was a split second ago.

By doing this, the complex "bouncing" equation suddenly starts to look like the "slow sliding" equation.
Then, by making those time steps even smaller, the "slow sliding" equation starts to look like the "frozen" equation.

It's like zooming in on a digital photo. At first, it looks like a smooth curve (the complex reality). But if you zoom in enough, you see it's actually made of tiny square pixels (the discrete steps). The authors prove that if you zoom out again, the pixels perfectly reconstruct the smooth curve.

3. The "Small Push" (Linearization)

Now, let's add the "Small Push." The authors assume the force you apply to the jelly is tiny.

The Big Idea: If you push the jelly only a microscopic amount, it doesn't matter if the jelly is made of complex, weird material. For that tiny movement, it behaves like a simple, perfect spring.
The Proof: The authors rigorously prove that as the "push" gets smaller and smaller (approaching zero), the complex, messy equations mathematically transform into the simple, linear equations everyone uses in engineering.

They show that this works at every level:

The complex "bouncing" model becomes the simple "bouncing" model.
The complex "sliding" model becomes the simple "sliding" model.
The complex "frozen" model becomes the simple "frozen" model.

4. The "Energy" Connection (The Glue)

How do they know this transformation is real and not just a lucky guess? They use a concept called $\Gamma$ -convergence (Gamma-convergence).

The Analogy: Think of the material's behavior as a landscape with hills and valleys.

The "Energy" of the system is like the height of the land. The material wants to roll down to the lowest valley (the most stable state).
The complex model has a very jagged, bumpy landscape.
The simple model has a smooth, bowl-shaped valley.

The authors prove that as the "push" gets smaller, the jagged, bumpy landscape smooths out perfectly into the bowl shape. Because the "landscape" (the energy) transforms correctly, the "ball" (the solution) rolling down it must also transform correctly.

Why Does This Matter?

In the real world, engineers often use the Simple Model (Linear) because it's easy to calculate on a computer. But they worry: "Is this accurate? What if the material is actually doing something weird?"

This paper says: "You are safe to use the simple model, provided the forces are small."

They have built a mathematical bridge that proves the simple model isn't just a "rough guess"—it is the rigorous, inevitable result of the complex reality when things are small. They showed that you can take the most complicated, realistic physics and, through a series of logical steps (zooming in on time, shrinking the force), arrive exactly at the simple equations we use every day.

In short: They took a tangled ball of yarn (the complex nonlinear world) and showed that if you pull on the ends gently enough, it unravels perfectly into a straight, simple string (the linear world), proving that the string was always there, hidden inside the tangle.

1. Problem Statement

The paper addresses the rigorous mathematical derivation of linearized viscoelastodynamics equations (specifically in Kelvin-Voigt rheology) from a nonlinear model of generalized standard materials.

Context: Nonlinear viscoelastodynamics describes the deformation $y(t, x)$ of a solid body under forces, governed by a balance of momentum involving inertial, elastic, and viscous forces.
Challenge: While linearization is standard in engineering for small strains, a rigorous derivation that preserves the dynamic nature (inertial effects) and the specific rheology (Kelvin-Voigt) from a fully nonlinear, frame-indifferent model has been an open challenge.
Scope: The authors consider a "nonsimple" material model (incorporating second-order gradients for regularization) to ensure the existence of weak solutions in the nonlinear setting, which is currently an open problem for simple materials.
Goal: To prove that as the strain parameter $\delta \to 0$ (representing vanishingly small loads and deformations), the solutions of the nonlinear dynamic system, the time-delayed (quasi-static approximation) system, and the fully discrete system all converge to the unique solution of the linearized dynamic system.

2. Methodology

The authors employ a multi-scale approach combining variational convergence ( $\Gamma$ -convergence) with time-discretization schemes and energy estimates.

A. The Approximation Scheme

The paper introduces a hierarchical approximation strategy to bridge the gap between fully dynamic, quasi-static, and static problems:

Time-Delayed Approximation (NL-TD): The inertial term $\varrho \partial_{tt} y$ is replaced by a difference quotient involving a time delay $h$ : $\varrho \frac{\partial_t y(t) - \partial_t y(t-h)}{h}$ . This transforms the hyperbolic problem into a sequence of quasi-static problems.
Fully Discrete Approximation (NL-disc): A second time scale $\tau$ is introduced, discretizing the time-delayed problem into a variational minimization problem (Minimizing Movements / Time-Incremental Minimization).
Linearization Limit ( $\delta \to 0$ ): The deformation is rescaled as $y = \text{id} + \delta u$ , where $\delta$ is a small parameter. The energy and dissipation functionals are rescaled accordingly to extract the linear limit.

B. Mathematical Tools

Generalized Standard Materials: The stress tensor is derived from an elastic energy functional $E$ and a dissipation functional $R$ . The model includes a "nonsimple" term (second gradient $\nabla^2 y$ ) to ensure compactness and existence of solutions.
$\Gamma$ -Convergence: Used to prove the convergence of the underlying energy and dissipation functionals as $\delta \to 0$ . This establishes that the minimizers of the nonlinear functionals converge to minimizers of the linearized functionals.
EDP-Convergence (Energy-Dissipation Principle): The authors utilize the concept of EDP-convergence (extended to include external forces) to show that the convergence of functionals implies the convergence of the associated gradient flow solutions.
Rigidity Estimates: They rely on the Friesecke-James-Müller rigidity estimate and generalized Korn inequalities to control the deformation gradients and ensure the displacement $u$ remains bounded in appropriate Sobolev spaces.
Minty Trick: Used to handle the convergence of the nonlinear stress terms (specifically the second-order hyperstress) in the weak formulation.

3. Key Contributions

Rigorous Derivation of Linearized Dynamics: The paper provides the first rigorous proof that the linearized Kelvin-Voigt viscoelastic equations are the $\Gamma$ -limit of the fully nonlinear dynamic equations under small strain assumptions, preserving the time-dependent frame indifference.
Commutativity of Limits: A central result is the demonstration that the linearization limit ( $\delta \to 0$ ) commutes with the approximation procedures (time-delay $h \to 0$ $h \to 0$ and time-discretization $\tau \to 0$ $τ \to 0$ ).
- One can linearize first and then discretize, or discretize first and then linearize; the resulting linearized dynamic solution is the same.
- This validates the use of linearized models in dynamic simulations, provided the loads and strains are small.
Handling of Inertial Terms: Unlike many previous works that focused on quasi-static or rate-independent systems, this paper explicitly handles the inertial term ( $\varrho \partial_{tt} u$ ) in the limit process, ensuring the resulting linear model is a hyperbolic system (wave equation with damping) rather than a parabolic or elliptic one.
Nonsimple Material Framework: By utilizing nonsimple materials (second-gradient elasticity), the authors bypass the current lack of existence theory for weak solutions in simple hyperelastic materials, allowing for a complete analysis of the nonlinear-to-linear transition.

4. Main Results

The paper establishes a comprehensive convergence diagram (visualized in the Introduction) covering three levels of approximation:

Nonlinear Discrete (NL-disc) $\to$ Linear Discrete (L-disc)
Nonlinear Time-Delayed (NL-TD) $\to$ Linear Time-Delayed (L-TD)
Nonlinear Dynamic (NL) $\to$ Linear Dynamic (L)

Specific Theorems:

Theorem 4.9: Proves the $\Gamma$ -convergence of the nonlinear discrete cost functional to the linearized discrete cost functional.
Proposition 6.9 & 6.10: Establishes the convergence of solutions from the nonlinear discrete problem to the linear discrete problem, and the simultaneous convergence as both $\delta \to 0$ and $\tau \to 0$ .
Theorem 7.4: Proves the existence of weak solutions for the fully nonlinear and linear dynamic problems via the limit $h \to 0$ of the time-delayed problems.
Theorem 7.8: Proves the direct convergence of nonlinear dynamic solutions to linear dynamic solutions as $\delta \to 0$ .
Theorem 8.5: Establishes EDP-convergence with forces, showing that the sequence of functionals $(E_\delta, R_\delta)$ converges to $(E_0, R_0)$ in a way that guarantees the convergence of the associated gradient flow solutions under external forces.

5. Significance

Theoretical Foundation: This work bridges a critical gap between nonlinear continuum mechanics and linearized elasticity. It provides a rigorous justification for using linearized models in dynamic scenarios, which is ubiquitous in engineering but often lacks a formal proof for complex viscoelastic materials.
Model Reduction: It validates the "Minimizing Movements" scheme as a robust tool for simulating dynamic viscoelasticity, showing that the discrete variational approach naturally leads to the correct linearized continuous equations.
Frame Indifference: The derivation strictly adheres to the principle of time-dependent frame indifference (Antman's sense), ensuring the physical consistency of the derived linearized model.
Future Applications: The methodology (combining $\Gamma$ -convergence, time-delay approximations, and EDP-convergence) provides a blueprint for analyzing other complex dynamic systems, such as thermoviscoelasticity or materials with microstructure, where linearization is desired but difficult to justify rigorously.

In summary, Benešová, Kampschulte, and Kružík successfully demonstrate that the linearized Kelvin-Voigt equations are the natural, rigorous limit of their nonlinear counterparts, even in the presence of inertia and external forces, provided the material is modeled with sufficient regularity (nonsimple materials).