Table of Contents
Fetching ...

Well-posedness of the Euler system of gas dynamics

Eduard Feireisl, Maria Lukacova-Medvidova

TL;DR

The paper addresses non-uniqueness in the Euler system for gas dynamics by proposing two principled selection procedures within the framework of dissipative measure-valued solutions. The two-step approach first selects entropy-admissible DMV solutions via maximal entropy production, then isolates a unique turbulent solution by minimizing the mean energy, with the energy defect vanishing at large times; a parallel one-step criterion based on a distance to equilibrium provides an alternative path. The authors establish well-posedness in a measurable, semigroup setting and prove key properties such as DiPerna maximality for Step 1, measurability and convergence results for Step 2, and the asymptotic vanishing of turbulent energy. They also extend the framework to a maximal-dissipativity criterion and discuss numerical implications, edge cases, and the mathematical structure (e.g., Moreau–Yosida regularisation, Mosco convergence). Overall, the work offers a rigorous route to a unique, physically consistent evolution for the Euler system under general initial data.

Abstract

We propose a new two-step selection criterion applicable to the dissipative measure--valued solutions of the Euler system of gas dynamics. The process consists of a successive maximisation of the entropy production rate and the total energy defect, i.e. maximisation of the turbulent energy. If the selected solution is a weak solution of the Euler system, then it is identified in the first step. Solutions selected in the second step are truly measure--valued maximising the energy defect. Accordingly, they are called turbulent solutions. The energy defect of turbulent solutions vanishes with growing time. The selected solutions depend in a Borel--measurable way on the initial data. In particular, they are almost continuously dependent on the initial data.

Well-posedness of the Euler system of gas dynamics

TL;DR

The paper addresses non-uniqueness in the Euler system for gas dynamics by proposing two principled selection procedures within the framework of dissipative measure-valued solutions. The two-step approach first selects entropy-admissible DMV solutions via maximal entropy production, then isolates a unique turbulent solution by minimizing the mean energy, with the energy defect vanishing at large times; a parallel one-step criterion based on a distance to equilibrium provides an alternative path. The authors establish well-posedness in a measurable, semigroup setting and prove key properties such as DiPerna maximality for Step 1, measurability and convergence results for Step 2, and the asymptotic vanishing of turbulent energy. They also extend the framework to a maximal-dissipativity criterion and discuss numerical implications, edge cases, and the mathematical structure (e.g., Moreau–Yosida regularisation, Mosco convergence). Overall, the work offers a rigorous route to a unique, physically consistent evolution for the Euler system under general initial data.

Abstract

We propose a new two-step selection criterion applicable to the dissipative measure--valued solutions of the Euler system of gas dynamics. The process consists of a successive maximisation of the entropy production rate and the total energy defect, i.e. maximisation of the turbulent energy. If the selected solution is a weak solution of the Euler system, then it is identified in the first step. Solutions selected in the second step are truly measure--valued maximising the energy defect. Accordingly, they are called turbulent solutions. The energy defect of turbulent solutions vanishes with growing time. The selected solutions depend in a Borel--measurable way on the initial data. In particular, they are almost continuously dependent on the initial data.

Paper Structure

This paper contains 15 sections, 7 theorems, 108 equations.

Key Result

Lemma 3.2

Let $\{ \mathcal{V}^i_{t,x} \}, \mathfrak{C}^i$ be two DMV solutions of the Euler system in $(0, T_i) \times \Omega$, with the initial data $(\varrho_0^i, \bm{m}_0^i, S^i_0, \mathcal{E}_0)$, $i=1,2$ respectively. Let be the associated dissipative solutions. Suppose that and Then the concatenated solution $(\{ \mathcal{V}_{t,x} \}, \mathfrak{C}) = (\{\mathcal{V}^1_{t,x}\} ,\mathfrak{C}_1) \cup_

Theorems & Definitions (14)

  • Definition 2.1: DMV solution
  • Remark 3.1
  • Lemma 3.2: Concatenation of DMV solutions
  • Proposition 3.3
  • proof
  • Definition 3.4: Absolute entropy maximiser
  • Theorem 3.5: Properties of solutions selected in Step 1
  • Lemma 4.1
  • proof
  • Remark 4.2
  • ...and 4 more