Abstract
We consider the Euler system of gas dynamics endowed with the incomplete \((e - \varrho - p)\) equation of state relating the internal energy e to the mass density \(\varrho \) and the pressure p. We show that any sufficiently smooth solution can be recovered as a vanishing viscosity-heat conductivity limit of the Navier–Stokes–Fourier system with a properly defined temperature. The result is unconditional in the case of the Navier type (slip) boundary conditions and extends to the no-slip condition for the velocity under some extra hypotheses of Kato’s type concerning the behavior of the fluid in the boundary layer.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
The Euler system describing the evolution of the density \(\varrho = \varrho (t,x)\), the velocity \(\mathbf{u}= \mathbf{u}(t,x)\), and the internal energy \(e = e(t,x)\) of a compressible inviscid fluid reads
The fluid is confined to a bounded domain \(\Omega \subset R^3\), with impermeable boundary,
The system (1.1) rewritten in terms of the phase variables \((\varrho , \mathbf{u}, e)\) is symmetric hyperbolic, see e.g. Benzoni-Gavage and Serre [4, Chapter 13, Section 13.2.2]. The problem is formally closed by prescribing a suitable equation of state (EOS). We consider a polytropic EOS
The equation of state (1.3) is incomplete, in particular, the (absolute) temperature \(\vartheta \) is not uniquely determined. Indeed Gibbs’ law asserts
where s is a new thermodynamic variable called entropy. Here \(D=\left( \frac{\partial }{\partial \varrho },\frac{\partial }{\partial \vartheta }\right) \). Plugging (1.3) in (1.4) we obtain a first order system that can be integrated yielding
and, in accordance with (1.3), (1.4),
for an arbitrary function P. Thus the absolute temperature \(\vartheta \) is determined by \(\varrho \) and e modulo the function P, see Cowperthwaite [6], Müller and Ruggeri [17], or [11, Chapters 2,3].
The Navier–Stokes–Fourier system describing the motion of a real viscous and heat conductive gas can be viewed as a viscous regularization of (1.1):
with the viscous stress \({\mathbb {S}}\) given by Newton’s rheological law
and the heat flux given by Fourier’s law
The second law of thermodynamics requires the entropy production rate
to be non-negative; whence the diffusion transport coefficients \({\widetilde{\mu }}\), \({\widetilde{\eta }}\), and \({\widetilde{\kappa }}\) must be non-negative. Note that, unlike in the Euler system (1.1), the knowledge of the temperature \(\vartheta \) is necessary to determine the entropy as well as the heat flux in (1.7). The internal energy e can be evaluated in terms of \(\varrho \), \(\vartheta \) through (1.6). Thus solutions of the associated Navier–Stokes–Fourier system (1.7), that may be seen as a viscous regularization of the Euler system (1.1), depend on the choice of P in (1.5).
We consider the vanishing dissipation limit of the Navier–Stokes–Fourier system, specifically, we rescale
Moreover, the existing mathematical theory of the Navier–Stokes–Fourier system (see [11]) is based on the augmentation of the pressure, and, accordingly, the internal energy and entropy, by the radiation component
The parameter a is very small and usually neglected in the real world applications. Consistently with (1.10), we therefore consider
Suppose that \(\gamma > 1\) is given and that the Euler system (1.1)–(1.3) admits a smooth (\(C^1\)) solution on a time interval [0, T]. Our goal is to identify the function P in (1.5) in such a way that any sequence of weak solutions to the Navier–Stokes–Fourier system (1.7)–(1.9) converges in the vanishing viscosity/radiation limit (1.10)–(1.12) to the solution of the Euler system in \((0,T) \times \Omega \). Moreover, we show that the convergence is unconditional, if the boundary layer is eliminated by the choice of the complete slip boundary conditions
where \(\mathbf{n}\) denotes the outer normal vector to \(\partial \Omega \). In the case of the no-slip boundary conditions
the convergence is conditioned by extra hypotheses of Kato’s type [15, 16] identified in the compressible setting by Sueur [21] and Wang and Zhu [22].
In comparison with the existing literature, notably [22], our result covers all admissible values of the adiabatic coefficient \(\gamma \) in (1.3) as well as general dependence of the transport coefficients on the temperature in the spirit of the existence theory developed in [11].
The paper is organized as follows. In Sect. 2, we recall the necessary preliminary material concerning the weak solutions to the Navier–Stokes–Fourier system including the relative energy inequality that represents a crucial tool in the analysis. Section 3 contains the main results. In Sect. 4 we show consistency of the vanishing viscosity approximation. Specifically, the viscous stress, the heat flux as well as the radiation components of the pressure, internal energy, and entropy along with the associated fluxes disappear in the regime specified in (1.10), (1.12). This process is “path dependent”, specifically certain relations concerning the asymptotic behaviour of \((\mu _n, \kappa _n, a_n)\) must be imposed in the spirit of [8]. The convergence towards the strong solution of the Euler system is shown in Sect. 5.
2 Preliminary Material
We recall the existing theory of weak solutions to the Navier–Stokes–Fourier system.
2.1 Mathematical Theory of the Closed System
We suppose the fluid is mechanically insulated as we stipulate either the complete slip (1.13) or the no-slip boundary condition (1.14). In view of our final objective, we require the fluid to be energetically isolated, specifically
The mathematical theory for closed systems relevant for future analysis was developed in [11]. Note that the extension to open systems is also available in the recent works [5, 12], see also the forthcoming monograph [9].
A suitable weak formulation of the Navier–Stokes–Fourier system augmented by the radiative terms proposed in [11] reads
see [11, Chapter 3]. Note that we anticipate the influence of thermal radiation represented by the extra terms \(p_R\), \(e_R\), and \(s_R\) in (2.2). In accordance with (1.12), these terms will vanish in the asymptotic limit. The energy balance appearing in the Euler system in (1.1) is replaced by the entropy inequality supplemented with the total energy balance in (2.2).
2.2 Transport Coefficients
In accordance with the molecular theory of gases (see e.g. Becker [1]), the transport coefficients depend on the temperature. Specifically, we assume that \({\widetilde{\mu }}\), \({\widetilde{\eta }}\), \({\widetilde{\kappa }}\) are continuously differentiable functions of \(\vartheta \) satisfying
for any \(\vartheta \ge 0\). Note that the cubic growth of \(\kappa \) is motivated by the presence of the radiation terms, see Oxenius [18].
2.3 Equation of State
A proper choice of the equation of state for the Navier–Stokes–Fourier system plays of course a crucial role in the present paper. Given \(\gamma > 1\), we have to identify the function P in (1.5). For \(p = p(\varrho , \vartheta )\), \(e = e(\varrho , \vartheta )\), we recall the hypothesis of thermodynamic stability
This imposes the following restrictions on P:
The following lemma shows existence of a suitable P.
Lemma 2.1
For all \({\underline{Z}}>0\) there exist functions \(P,S\in C^1[0,\infty )\) with properties (1.6), (2.5) and such that
Moreover P, S satisfy
Note that according to (2.10), S from Lemma 2.1 is in accordance with the Third law of thermodynamics, namely
Proof
Let us first consider the case \({\underline{Z}}=1\). Set \(P,S\in C^1[0,\infty )\)
and
It is then straightforward to check (1.6), (2.5), (2.6)–(2.10).
Let us now look at \({\underline{Z}}\ne 1\). We define the P, S constructed above as \(P_1,S_1\) and set
Again straightforward computations show that the properties (1.6), (2.5), (2.6)–(2.10) follow from the corresponding property of \(P_1,S_1\). \(\square \)
Note that for \(Z\in [0,{\underline{Z}}]\), according to (2.6) and (1.5) we simply obtain the Boyle-Mariotte law
Hence the temperature for the Euler system (1.1) endowed with the incomplete EOS (1.3) can be recovered by choosing \({\underline{Z}}\) in Lemma 2.1 appropriately, see Sect. 4.1 for details.
2.4 Relative Energy
The relative energy for the Navier–Stokes–Fourier system may be seen as a counterpart of Dafermos’ relative entropy for the (hyperbolic) Euler system, see [7]. Given a trio of “test functions”
the relative energy reads
where
is the ballistic free energy. In the context of the system (2.2) perturbed by the radiation terms, we have
The relative energy augmented by the radiation component will be denoted \(E_a\). We also introduce the standard energy
The following result was proved in [10]: Suppose that:
-
\((\varrho , \vartheta , \mathbf{u})\) is a weak solution to the Navier–Stokes–Fourier system (2.2) in \((0,T) \times \Omega \) with the no-flux boundary conditions (2.1) and either the complete slip boundary conditions (1.13) or the no-slip boundary condition (1.14).
-
\((r, \Theta , \mathbf{U})\) is a trio of continuously differentiable test functions,
$$\begin{aligned} r> 0,\ \Theta > 0 \ \text{ in }\ [0,T] \times \overline{\Omega }, \end{aligned}$$where \(\mathbf{U}\) satisfies either the impermeability boundary condition
$$\begin{aligned} \mathbf{U} \cdot \mathbf{n}|_{\partial \Omega } = 0, \end{aligned}$$or the no-slip boundary condition
$$\begin{aligned} \mathbf{U}|_{\partial \Omega } = 0. \end{aligned}$$
Then the relative energy inequality
holds for a.a. \(\tau \in (0,T)\).
Finally, we recall the fundamental properties of the relative energy that follow from the hypothesis of thermodynamic stability (2.4). In accordance with hypothesis (2.11), fix
and define
where
Then
where the constants depend on \({\underline{\varrho }}\), \(\overline{\varrho }\), \({\underline{\vartheta }}\), and \(\overline{\vartheta }\), see e.g. [11] for details. As a consequence of the hypothesis of thermodynamic stability (2.4), the relative energy expressed in terms of the conservative entropy variables \((\varrho , \mathbf{m} = \varrho \mathbf{u}, {\mathcal {S}} = \varrho s )\) is a strictly convex function and represents the so-called Bregman distance between \((\varrho , \mathbf{m}, {\mathcal {S}})\) and \((r, r\mathbf{U}, r s(r, \Theta ) )\), see e.g. [12]. Note carefully that the relative energy \(E_a\) associated to the Navier–Stokes–Fourier system (2.2) is augmented by the radiation component
3 Main Results
We state the main results in the physically relevant case \(\Omega \subset R^3\). We consider three vanishing parameters in the asymptotic limit: the viscosity coefficient \(\mu _n\), the heat conductivity coefficient \(\kappa _n\), and the radiation parameter \(a_n\), cf. (1.10), (1.12).
3.1 Unconditional Convergence in the Absence of Boundary Layer
We start with the Navier–Stokes–Fourier system (2.2), with the complete slip boundary conditions (1.13), and the no-flux boundary condition (2.1).
Theorem 3.1
(Unconditional convergence). Let \(\Omega \subset R^3\) be a bounded Lipschitz domain. Suppose that the Euler system (1.1)–(1.3), with \(\gamma > 1\), admits a strong solution
satisfying
Then there exists a (p-\(\varrho \)-\(\vartheta \)) EOS \(p = p(\varrho , \vartheta )\) that complies with Gibbs’ relation (1.4) as well as the hypothesis of thermodynamic stability (2.4), with the associated internal energy EOS \(e = e(\varrho , \vartheta )\) and entropy \(s = s(\varrho , \vartheta )\) determined through (1.6), such that the following holds:
Let \((\varrho _n, \vartheta _n, \mathbf{u}_n)_{n=1}^\infty \) be a sequence of weak solutions to the Navier–Stokes–Fourier system (2.2), with the complete slip boundary condition (1.13), and the no-flux boundary conditions (2.1), in the vanishing dissipation/radiation regime:
where \(\alpha \in [\frac{1}{3}, 1]\) is the exponent in hypothesis (2.3). In addition, suppose that the initial data
converge strongly to those of the Euler system, specifically,
Then
Remark 3.2
The reader may consult [11, Chapter 3] for the exact definition of a weak solution of the Navier–Stokes–Fourier system emanating from the initial data \((\varrho _{n,0}, \vartheta _{n,0}, \mathbf{u}_{n,0})\).
Remark 3.3
We strongly point out that Theorem 3.1does not contain any claim concerning the existence of weak solutions for the Navier–Stokes–Fourier system. The existence is known only in some particular cases: \(\gamma \ge \frac{5}{3}\), \(\alpha \in [\frac{2}{5}, 1]\), see [11, Chapter 3, Theorem 3.1], and \(\gamma > \frac{3}{2}\), \(\alpha = 1\), see Jesslé, Jin, and Novotný [14, Theorem 2.1]. The best known results for the planar flows were obtained recently by Pokorný and Skříšovský [19].
Local in time existence of smooth solutions to the Euler system was established by Schochet [20].
3.2 Conditional Result: Viscous Boundary Layer
The no-slip boundary condition (1.14) imposed on the viscous flow cannot be retained for the limit Euler system and the well known problem of viscous boundary layer appears. We report conditional results à la Kato in the spirit of Sueur [21] and Wang, Zhu [22]. Let
Any vector field \(\mathbf{w}\) can be decomposed into its normal and tangential component with respect to \(\partial \Omega \):
Note that \(|\nabla _x\mathrm{dist}[x, \partial \Omega ]|=1\), see Sect. 5.1.
We start with a result inspired by Sueur [21].
Theorem 3.4
(Conditional convergence, gradient criterion). Let \(\Omega \subset R^3\) be a bounded domain of class \(C^{2 + \nu }\). Suppose that the Euler system (1.1)–(1.3), with \(\gamma > 1\), admits a strong solution
satisfying
Then there exists a (p-\(\varrho \)-\(\vartheta \)) EOS \(p = p(\varrho , \vartheta )\) that complies with Gibbs’ relation (1.4) as well as the hypothesis of thermodynamic stability (2.4), with the associated internal energy EOS \(e = e(\varrho , \vartheta )\) and entropy \(s = s(\varrho , \vartheta )\) determined through (1.6), such that the following holds:
Let \((\varrho _n, \vartheta _n, \mathbf{u}_n)_{n=1}^\infty \) be a sequence of weak solutions to the Navier–Stokes–Fourier system (2.2), with the no-slip boundary condition (1.14), and the no-flux boundary conditions (2.1), in the vanishing dissipation/radiation regime:
where \(\alpha \in [\frac{1}{3}, 1]\) is the exponent in hypothesis (2.3). In addition, suppose that the initial data
converge strongly to those of the Euler system, specifically,
Finally, supposeFootnote 1
as \(n \rightarrow \infty \).
Then
Finally, we state a conditional result inspired by Wang and Zhu [22].
Theorem 3.5
(Conditional convergence). Let \(\Omega \subset R^3\) be a bounded domain of class \(C^{2 + \nu }\). Suppose that the Euler system (1.1)–(1.3), with \(\gamma > 1\), admits a strong solution
satisfying
Then there exists a (p-\(\varrho \)-\(\vartheta \)) EOS \(p = p(\varrho , \vartheta )\) that complies with Gibbs’ relation (1.4) as well as the hypothesis of thermodynamic stability (2.4), with the associated internal energy EOS \(e = e(\varrho , \vartheta )\) and entropy \(s = s(\varrho , \vartheta )\) determined through (1.6), such that the following holds:
Let \((\varrho _n, \vartheta _n, \mathbf{u}_n)_{n=1}^\infty \) be a sequence of weak solutions to the Navier–Stokes–Fourier system (2.2), with the no-slip boundary condition (1.14), and the no-flux boundary conditions (2.1), in the vanishing dissipation/radiation regime:
where \(\alpha \in [\frac{1}{3}, 1]\) is the exponent in hypothesis (2.3). In addition, suppose that the initial data
converge strongly to those of the Euler system, specifically,
Finally, suppose there is a sequence \(\delta _n \rightarrow 0\) such that
uniformly for \(n \rightarrow \infty \).
Then
Hypothesis (3.5) may seem awkward and much stronger than its counterpart by Wang and Zhu [22], where only the case \(\alpha =1\) is studied. In order to compare our result with [22], we are able to modify Theorem 3.5 for \(\alpha =1\) and obtain the following.
Theorem 3.6
(Conditional convergence, \(\alpha = 1\)). Let \(\Omega \subset R^3\) be a bounded domain of class \(C^{2 + \nu }\). Suppose that the Euler system (1.1)–(1.3), with \(\gamma > 1\), admits a strong solution
satisfying
Then there exists a (p-\(\varrho \)-\(\vartheta \)) EOS \(p = p(\varrho , \vartheta )\) that complies with Gibbs’ relation (1.4) as well as the hypothesis of thermodynamic stability (2.4), with the associated internal energy EOS \(e = e(\varrho , \vartheta )\) and entropy \(s = s(\varrho , \vartheta )\) determined through (1.6), such that the following holds:
Let \((\varrho _n, \vartheta _n, \mathbf{u}_n)_{n=1}^\infty \) be a sequence of weak solutions to the Navier–Stokes–Fourier system (2.2), with the no-slip boundary condition (1.14), and the no-flux boundary conditions (2.1), in the vanishing dissipation/radiation regime with \(\alpha = 1\):
In addition, suppose that the initial data
converge strongly to those of the Euler system, specifically,
Finally, suppose there is a sequence \(\delta _n \rightarrow 0\) such that
as \(n \rightarrow \infty \).
Then
Remark 3.7
Indeed hypothesis (3.6) and the assumptions in Wang and Zhu [22] are similar, though not equivalent. Note furthermore, that Wang and Zhu alternatively consider an analogous assumption on the tangential component of \(\mathbf{u}_n\) instead of the normal component. In this paper we do not pursue anything of that kind.
The rest of the paper is devoted to the proof of the above results.
4 Consistency of the Vanishing Dissipation/Radiation Approximation
As a preliminary step, we show consistency of the vanishing dissipation/radiation approximation.
4.1 Temperature for the Euler System
First we introduce the temperature \(\vartheta _E\) associated to the limit system. Without loss of generality, we may fix the constants \({\underline{\varrho }}\), \(\overline{\varrho }\) in (3.2) so that
Next, in accordance with the hypotheses of Theorem 3.1,
for certain constants \({\underline{e}}\), \(\overline{e}\). Let us set
and apply Lemma 2.1 to obtain suitable functions P, S. Furthermore we define
Note that \(\vartheta _E>(\gamma -1){\underline{e}}\) and hence
By virtue of (2.6), we have
Moreover, without loss of generality, we may suppose
with the same constants \({\underline{\vartheta }}\), \(\overline{\vartheta }\) as in (3.2). From this moment on, the pressure law is fixed.
As p, e, and s comply with Gibbs’ relation, the smooth solution of the Euler system conserves the entropy:
where s is given by (1.6).
4.2 Consistency
The Navier–Stokes–Fourier system (2.2) may be viewed as a singular perturbation of the Euler system with the extra “error” terms
We say that the approximation of the Euler system by the Navier–Stokes–Fourier system is consistent, if the above “error” terms vanish in the asymptotic limit \(n \rightarrow 0\). As a matter of fact, we need a milder form of consistency compatible with the relative energy inequality. More specifically, it is sufficient to control the “errors” by the dissipation term
and the total energy
For each error term \(E^i_n\), \(i=1,\dots , 6\) specified in (4.5) and \(\varepsilon > 0\), we have to find \(c(\varepsilon )\) such that
Obviously, \(E^1_n = p_R\), \(E^6_n = \varrho e_R\), and \(E^3_n = \varrho s_R\) satisfy (4.6) (with \(\varepsilon = 0\)), it remains to handle the viscous stress, the heat flux and the entropy convective flux term.
Moreover, we recall some basic estimates that follow directly from the hypotheses (1.6), (2.5), (2.7)–(2.10):
4.2.1 Viscous Stress Consistency
By virtue of hypothesis (2.3),
where the last inequality follows from hypothesis (3.1) and the simple fact that \(x^{1+\alpha }\le c (x^4 + 1)\). Thus we obtain the desired estimate (4.6). The bulk viscosity term can be handled in a similar fashion.
4.2.2 Heat Flux Consistency
Similarly to the preceding part,
whence (4.6) follows from hypothesis (3.1).
4.2.3 Radiation Entropy Convective Flux Consistency
To close the circle of consistence estimates, we have to handle the integral
that corresponds to the radiation entropy convective flux.
By virtue of Sobolev embedding theorem,
and, by a generalized Korn–Poincaré inequality [11, Theorem 11.23],
Another application of Hölder’s inequality yields
Consequently,
Finally, by virtue of hypothesis (3.1)
The rightmost integral in (4.9) can be handled in a similar fashion. Since \(\alpha \in [\frac{1}{3},1]\), we have in particular \(0 \le \alpha \le 3\) and the desired conclusion (4.6) follows from the boundedness of the total energy.
5 Convergence
The proof of convergence consists of using the strong solution \((\varrho _E, \vartheta _E, \mathbf{u}_E)\) of the Euler system as the test functions \(r = \varrho _E\), \(\Theta = \vartheta _E\), \( \mathbf{U}= \mathbf{u}_E\) in the relative energy inequality (2.13). This can be done in a direct manner in the case of the complete slip boundary conditions (1.13) , whereas the velocity \(\mathbf{u}_E\) must be modified to comply with the homogeneous Dirichlet boundary conditions in the case of no-slip (1.14). We focus on the latter case as the proof in the case of the complete slip boundary conditions can be performed in a way similar to [8].
5.1 Velocity Regularization
If the solutions of the Navier–Stokes–Fourier system satisfy the no-slip boundary conditions, the velocity \(\mathbf{u}_E\) is not eligible for the relative energy inequality (2.13) as its tangential component may not vanish on \(\partial \Omega \). Instead we consider
where the perturbation \(\mathbf{v}_\delta \) is given as
where
and
If \(\partial \Omega \) is of class \(C^k\), \(k \ge 2\), then \(\mathrm{dist}[x, \partial \Omega ] \in C^k(\Omega _\delta )\) for any \(0< \delta < \delta _0\), and
see Foote [13].
5.2 Application of the Relative Energy Inequality
As \(\mathbf{U} = \mathbf{u}_E - \mathbf{v}_\delta \) vanishes on \(\partial \Omega \), the trio \((r = \varrho _E, \mathbf{U} = \mathbf{u}_E - \mathbf{v}_\delta , \Theta = \vartheta _E)\) can be used as test functions in the relative energy inequality (2.13). Recall that at this stage we have the following vanishing parameters: \(\mu _n\), \(\kappa _n\), \(a_n\), and \(\delta = \delta _n\).
We have
Seeing that
we may infer that
The first rather straightforward observation is that, under hypothesis (3.2) concerning the initial data,
Consequently, we can write (2.13) in the form
which holds for a.a. \(\tau \in (0,T)\), where h denotes a generic sequence,
Our goal is to show
by means of a Gronwall type argument.
5.3 Integrals Controlled by the Consistency Estimates
Evoking the bounds obtained in Sect. 4.2 we get
where, by virtue of (4.8),
Using the consistency estimates of Sect. 4.2, we can handle other integrals containing vanishing parameters. Accordingly, the inequality (5.6) simplifies to
Moreover, as \(\mathbf{u}_E \cdot \mathbf{n}|_{\partial \Omega } = 0\),
and, consequently,
Finally, as \((\varrho _E, \vartheta _E, \mathbf{u}_E)\) solves the Euler system,
In addition, it is easy to check that
Consequently,
as both \((\varrho _n )_{n \ge 0}\) and \((\varrho _n \mathbf{u}_n)_{n \ge 0}\) are equi-integrable in \((0,T) \times \Omega \). Thus (5.9) reduces to
5.4 Integrals Independent of the Boundary Layer
Now, we estimate the integrals on the right-hand side of (5.11) that are independent of \(\mathbf{v}_\delta \). First, by virtue of (4.7),
We point out that this step depends in an essential way on the fact that s satisfies the Third law of thermodynamics.
Next, we recall two identities that follow from the specific form of EOS (1.3), (1.4), namely
Consequently, we get
Finally, we use the identity
Plugging (5.14) into (5.13) yields the desired estimate. Thus (5.11) reduces to
Note that inequality (5.15) almost completes the proof of Theorem 3.1, where we may take \(\mathbf{v}_\delta = 0\). It only remains to show the desired strong convergence claimed in (3.3). This will be done in Sect. 5.6.
However, in order to prove Theorems 3.4–3.6, where \(\mathbf{v}_\delta \ne 0\), one has to estimate the first two integrals on the right-hand side of (5.15), which is carried out in the following Sect. 5.5.
5.5 Boundary Layer
It remains to control the first two integrals on the right-hand side of (5.15) that represent the effect of the boundary layer.
5.5.1 Viscous Stress
Similarly to Sect. 4.2, we have
where
Consequently, when proving Theorems 3.5 and 3.6, the desired estimate follows from hypothesis (3.5) and (3.6), respectively. Note that this type of estimates forces us to consider the thickness \(\delta \) of the boundary layer asymptotically larger than \(\mu \),
Alternatively, in order to show Theorem 3.4 and following Sueur [21], we have (3.4), meaning
Setting \(\mu _n \approx \delta _n\), we get
5.5.2 Convective Term
Finally, we consider
Recall that
Similarly, for a scalar function F, we decompose
In accordance with the definition of \(\mathbf{v}_\delta \), we get
Now,
where, by virtue of (5.17),
In view of (5.4), \((\varrho _n)_{n\ge 0}\), \((\varrho _n \mathbf{u}_n )_{n \ge 0}\) are equi-integrable; whence
uniformly in n.
Thus it remains to handle the integral
Let us first look at Theorem 3.5. By Hölder’s inequality and (5.17),
where \(\frac{24}{7 - 3\alpha }\) is the critical exponent in the Sobolev–Poincaré inequality
As \(\mathbf{u}_n|_{\partial \Omega } = 0\), Korn’s inequality yields
Note that the constants are independent of \(\delta \) as \(\mathbf{u}_n\) can be extended to be zero outside \(\Omega \).
Thus going back to (5.18) we deduce
in accordance with hypothesis (3.5).
Next, in view of Theorem 3.6 we consider \(\alpha = 1\) and replace the critical exponent \(\frac{24}{17 + 3 \alpha }\) in (5.18) by the \(L^2\)-norm. Consequently,
Now, replacing (5.19) by Hardy–Sobolev inequality, we gain the multiplicative factor \(\delta \),
Thus the final inequality reads
in accordance with (3.6).
Finally, in order to show Theorem 3.4, one has to estimate the left-hand side of (5.18) using hypothesis (3.4). This works exactly as in Sueur [21].
5.6 Strong Convergence
We have established the convergence
uniformly for a.a. \(\tau \in (0,T)\). This obviously yields
In addition, as the energy of the initial data converges and both Euler and the Navier–Stokes–Fourier system conserve energy, we get
This yields the desired strong convergence claimed in (3.3).
Notes
Note, that we use the index n both for the sequence and the normal component. Throughout this paper \(\mathbf{u}_n\) denotes the nth element of the sequence \((\mathbf{u}_n)_{n=1}^{\infty }\) and \((\mathbf{u}_n)_n\) its normal component.
References
Becker, E.: Gasdynamik. Teubner-Verlag, Stuttgart (1966)
Belgiorno, F.: Notes on the third law of thermodynamics, i. J. Phys. A 36, 8165–8193 (2003)
Belgiorno, F.: Notes on the third law of thermodynamics, ii. J. Phys. A 36, 8195–8221 (2003)
Benzoni-Gavage, S., Serre, D.: Multidimensional Hyperbolic Partial Differential Equations. First Order Systems and Applications. Oxford Mathematical Monographs. The Clarendon Press, Oxford University Press, Oxford (2007)
Chaudhuri, N., Feireisl, E.: Navier–Stokes–Fourier system with Dirichlet boundary conditions. arxiv preprint No. arXiv:2106.05315 (2021)
Cowperthwaite, M.: Relations between incomplete equations of state. J. Frank. Inst. 287, 379–387 (1969)
Dafermos, C.M.: The second law of thermodynamics and stability. Arch. Rational Mech. Anal. 70, 167–179 (1979)
Feireisl, E.: Vanishing dissipation limit for the Navier–Stokes–Fourier system. Commun. Math. Sci. 14(6), 1535–1551 (2016)
Feireisl, E., Novotný, A.: Mathematics of Open Fluid Systems. Birkhäuser-Verlag, Basel (2020)
Feireisl, E., Novotný, A.: Weak-strong uniqueness property for the full Navier–Stokes–Fourier system. Arch. Rational Mech. Anal. 204, 683–706 (2012)
Feireisl, E., Novotný, A.: Singular Limits in Thermodynamics of Viscous Fluids. Advances in Mathematical Fluid Mechanics, 2nd ed. Birkhäuser/Springer, Cham (2017)
Feireisl, E., Novotný, A.: Navier–Stokes–Fourier system with general boundary conditions. Commun. Math. Phys. 386(2), 975–1010 (2021)
Foote, R.L.: Regularity of the distance function. Proc. Am. Math. Soc. 92, 153–155 (1984)
Jesslé, D., Jin, B.J., Novotný, A.: Navier–Stokes–Fourier system on unbounded domains: weak solutions, relative entropies, weak-strong uniqueness. SIAM J. Math. Anal. 45(3), 1907–1951 (2013)
Kato, T.: Remarks on the zero viscosity limit for nonstationary Navier–Stokes flows with boundary. In: Chern, S.S. (Ed.) Seminar on PDE’s. Springer, New York (1984)
Kato, T.: Remarks on zero viscosity limit for nonstationary Navier–Stokes flows with boundary. In: Seminar on Nonlinear Partial Differential Equations (Berkeley, California, 1983), volume 2 of Mathematical Sciences Research Institute Publications, pp. 85–98. Springer, New York (1984)
Müller, I., Ruggeri, T.: Rational Extended Thermodynamics. Springer Tracts in Natural Philosophy vol. 37. Springer, Heidelberg (1998)
Oxenius, J.: Kinetic Theory of Particles and Photons. Springer, Berlin (1986)
Pokorný, M., Skříšovský, E.: Weak solutions for compressible Navier–Stokes–Fourier system in two space dimensions with adiabatic exponent almost one. Acta Appl. Math. 172:Paper No. 1, 31 (2021)
Schochet, S.: The compressible Euler equations in a bounded domain: existence of solutions and the incompressible limit. Commun. Math. Phys. 104, 49–75 (1986)
Sueur, F.: On the inviscid limit for the compressible Navier–Stokes system in an impermeable bounded domain. J. Math. Fluid Mech. 16(1), 163–178 (2014)
Wang, Y.-G., Zhu, S.-Y.: On the vanishing dissipation limit for the full Navier–Stokes–Fourier system with non-slip condition. J. Math. Fluid Mech. 20(2), 393–419 (2018)
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no conflict of interest.
Additional information
Communicated by T. Ozawa.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This article is part of the topical collection “Yoshihiro Shibata” edited by Tohru Ozawa.
The research of Eduard Feireisl leading to these results has received funding from the Czech Sciences Foundation (GAČR), Grant Agreement 21-02411S. The Institute of Mathematics of the Academy of Sciences of the Czech Republic is supported by RVO:67985840.
Simon Markfelder acknowledges financial support by the Alexander von Humboldt Foundation.
Rights and permissions
About this article
Cite this article
Feireisl, E., Klingenberg, C. & Markfelder, S. Euler System with a Polytropic Equation of State as a Vanishing Viscosity Limit. J. Math. Fluid Mech. 24, 67 (2022). https://doi.org/10.1007/s00021-022-00690-7
Accepted:
Published:
DOI: https://doi.org/10.1007/s00021-022-00690-7