Abstract
We consider smooth solutions to a relaxed Euler system with Oldroyd-type constitutive laws. This system is derived from the one-dimensional compressible full Navier-Stokes equations for a Newtonian fluid by using the Cattaneo–Christov model and the Oldroyd-B model. In a neighborhood of equilibrium states, we construct an explicit symmetrizer and show that the system is symmetrizable hyperbolic with partial dissipation. Moreover, by establishing uniform estimates with respect to the relaxation times, we prove the uniform global existence of smooth solutions and the global-in-time convergence of the system towards the full Navier–Stokes equations.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
Euler and Navier-Stokes equations are two important models in fluid dynamics. There are rich literatures on the mathematical analysis around these equations. We refer to [11, 23,24,25,26] and references therein for mathematical results. There are deep relationships between Euler and Navier-Stokes equations. It is well known that the Euler equations can be derived from Navier-Stokes equations through the vanishing viscosity limit. Meanwhile, the Navier-Stokes equations can also be approximated by first-order partial differential equations using different kinds of constitutive laws for non-Newtonian fluids. These approximate equations are referred to as relaxed Euler systems or hyperbolic Navier-Stokes equations, see for instance [10, 16, 30, 33, 34].
In the paper, we study the global-in-time convergence from relaxed Euler-type equations with Oldroyd’s constitutive laws to compressible full (non-isentropic) Navier-Stokes equations by letting relaxation times tend to zero. Let \(t\ge 0\) be the time variable and \(x=(x_1,\cdots , x_d)\in {\mathbb {R}}^d\) be the space variable. The compressible full Navier-Stokes equations are of the form
in \({\mathbb {R}}^+\times {\mathbb {R}}^d\), where \(\rho >0\) is the density, \(u=(u_1, \cdots , u_d)^\top \in {\mathbb {R}}^d\) is the velocity, \(\theta \) is the temperature, p is the pressure function, \(q\in {\mathbb {R}}^d\) is the heat flux, \(\tau \) is the stress tensor and \(E=e+\dfrac{1}{2}|u|^2\) is the total energy per unit mass with e being the specific internal energy. The symbols \(\top \) and \(\otimes \) represent the transpose and the tensor product, respectively. In (1.1), \(\rho \), u and \(\theta \) are independent variables, e and p are functions of \((\rho ,\theta )\). In particular, for the ideal fluid, we have
which satisfy the thermodynamic equation
where R and \(c_v\) are positive constants, and \(e_\rho =\frac{\partial e}{\partial \rho }\) etc. Generally speaking, the heat flux q satisfies the following Fourier’s law
where \(\kappa >0\) is the heat conduction constant. For Newtonian viscous fluids, the stress tensor \(\tau \) takes the form
with
Here \(I_d\) denotes the unit matrix of order d. The parameters \(\mu >0\) and \(\lambda >0\) are the shear and bulk viscosity coefficients, respectively, which are all assumed to be constants.
For the full Navier-Stokes equations (1.1) with constitutive laws (1.4) and (1.5), the construction of the corresponding relaxed Euler systems depends on the way how to decompose the second-order derivatives \(\mathrm {div}\tau , \mathrm {div}(u\tau )\) and \(\mathrm {div}q\) into first-order derivative terms. Clearly, there are lots of ways to do it, among which the most natural one is to replace (1.4) and (1.5) by the following Maxwell’s constitutive laws [28]
where \(\varepsilon _1>0\) and \(\varepsilon _2>0\) are the relaxation times. Equations in (1.1) together with (1.6)-(1.7) form a relaxed Euler system. Formally, letting \((\varepsilon _1,\varepsilon _2)\rightarrow 0\) recovers the Navier-Stokes equations (1.1) with (1.4) and (1.5). This idea is not recent. It dates back to 1860s, see for instance [3, 4, 28]. These approximations have not only the mathematical sense but also physical interpretations. Relation (1.6), also known as the Cattaneo’s law, gives rise to heat waves with finite propagation speed. Relation (1.7) describes motions of viscoelastic fluids. The laws are combinations of the Newtonian’s law of viscosity and the Hooke’s law of elasticity.
The existence of smooth solutions to system (1.1) with Maxwell’s constitutive laws (1.6) and (1.7) and their convergence to the classical non-isentropic Navier-Stokes equations (1.1) with (1.4) and (1.5) have been studied in previous works. In the case where \(\varepsilon _1=0\), the authors of [15] proved the global existence of smooth solutions near constant equilibrium states for fixed \(\varepsilon _2>0\) and the local-in-time convergence towards the Navier-Stokes equations as \(\varepsilon _2\rightarrow 0\). Similar results are obtained in [14] in the case where \(\varepsilon _2=0\) and \(\varepsilon _1>0\). In these results, only one of the constitutive laws within (1.6) and (1.7) is used. Hence the systems studied in [14, 15] are of mixed hyperbolic-parabolic type in the sense of Shizuta-Kawashima [19, 39]. The local-in-time convergence is based on the error estimates between the original system and the limiting system. For the isentropic Navier-Stokes equations with constitutive law (1.7), the author of [42] obtained the local existence and the local convergence to the classical isentropic Navier-Stokes equations under condition \(\text {tr}(\tau )=0\), where \(\text {tr}(\tau )\) means the trace of matrix \(\tau \). In [30], the first author of the present paper constructed approximate systems with vector variables instead of tensor variables by using Hurwitz-Radon matrices in both compressible and incompressible cases. He proved the uniform (with respect to \(\varepsilon _1\) and \(\varepsilon _2\)) global existence of smooth solutions near constant equilibrium state and the global-in-time convergence of the systems towards classical isentropic Navier-Stokes equations. He also obtained similar results for the isentropic Navier-Stokes equations with constitutive law (1.7) without condition \(\text {tr}(\tau )=0\). For the approximation of incompressible isentropic Navier-Stokes equations with constitutive law (1.7), we also refer to [10, 33, 34, 37, 38].
However, these two constitutive laws (1.6) and (1.7) have drawbacks as they do not ensure Galilean invariance. In other words, these laws lead to paradoxical evolution of thermal waves in a moving frame, see [7]. To overcome it, the Oldroyd’s upper-convected time derivative (or simply Oldroyd derivative) should be considered. In this paper, we consider the following two constitutive laws. The first one is the Cattaneo-Christov model introduced in [6],
in which the terms on the left-hand side are the Oldroyd derivative. It is proved in [6] that the constitutive law (1.8) is Galilean invariant. The second one is the following Oldroyd-B model for the tensor variable \(\tau \) (see for instance [2, 29, 35, 36] and the references therein)
where
Hence, the relaxed Euler system for (1.1) with constitutive laws (1.8) and (1.9) is of the form
in \({\mathbb {R}}^+\times {\mathbb {R}}^d\).
System (1.10) is very complicated. One can observe that it contains \(d^2+2d+2\) equations for fluids in space \({\mathbb {R}}^d\). So far the symmetrizable hyperbolicity for (1.10) is unknown in cases \(d\ge 2\). This makes it hard to establish the existence results. See classical theories [18, 21, 24]. Moreover, when considering the constitutive laws (1.8) and (1.9) at the same time, there are no apparent dissipative structures for \(\nabla u\) and \(\nabla \theta \) due to the loss of the elliptic structures for u and \(\theta \). This is different from the situation for the systems treated in [14, 15].
In a recent paper [16], the authors considered (1.10) in one space dimension. In this case, system (1.10) is reduced to the following form
in \({\mathbb {R}}^+\times {\mathbb {R}}\), with the initial data
where \(\varepsilon =(\varepsilon _1,\varepsilon _2)\). In the limit as \(\varepsilon \rightarrow 0\), we have formally
Substituting these relations into (1.11), we recover the following one-dimensional non-isentropic Navier-Stokes equations
in \({\mathbb {R}}^+\times {\mathbb {R}}\). Equations (1.13) have been widely studied. We refer to [13, 17, 20, 27] for the global existence of smooth solutions. See also [11, 23] for related topics and references therein. In (1.11), \(\rho \), u, \(\theta \), q and \(\tau \) are independent variables, and both the internal energy e and the pressure function p are functions of \((\rho ,\theta ,q,\tau )\). In [16], the authors used the following state equations for e and p :
Equation (1.14) is based on the results in [8], where the authors rigorously proved that the constitutive law (1.8) is consistent with the second law of thermodynamics if and only if the dependence of e on q is quadratic. Similarly, the quadratic dependence of e on \(\tau \) also implies the compatibility with the second law of thermodynamics. The choice of equation (1.15) makes it consistent with (1.3) and the state equations (1.14)-(1.15) yield formally those for the ideal fluid (1.2) as \(\varepsilon \rightarrow 0\). For more explanations, see [5, 9, 16, 41].
Let \(V_e=(1,0,1,0,0)^\top \) be an equilibrium state for (1.11). Due to the explicit expressions (1.14) and (1.15), the authors of [16] constructed a strictly convex entropy for system (1.11), see Lemma 3.1 in [16] or Lemma 3.1 below. Based on this, they established the global existence of smooth solutions near \(V_e\) for fixed \(\varepsilon _1>0\) and \(\varepsilon _2>0\), and the local convergence of the system towards the classical full Navier-Stokes equations as \(\varepsilon _1=\varepsilon _2\rightarrow 0\). However, the convergence for large time has not been investigated.
The purpose of this paper is to study the global convergence of system (1.11) with state equations (1.14) and (1.15). The main results of the paper are contained in the following two theorems. The first theorem shows the uniform global existence of smooth solutions to the Cauchy problem (1.11)-(1.12) near \(V_e\). The second one concerns the global convergence of the solution of (1.11) to that of the one-dimensional full Navier-Stokes equations (1.13) for the ideal fluid (1.2) as \(\varepsilon \rightarrow 0\). Remark that in these results condition \(\varepsilon _1=\varepsilon _2\) is not needed and system (1.11) is not included in the class of systems studied in [22, 31, 32, 43]. The proof of the results is based on uniform estimates with respect to the time and the relaxation parameters. We use the strictly convex entropy given in [16] for the \(L^2\) estimate. A key step is to find an appropriate symmetrizer of system (1.11) for higher-order estimates. It is well-known that the second-order derivative of a strictly convex entropy provides a symmetrizer for a system of conservation laws [1, 12]. However, this result cannot be applied to (1.11) because it is a non-conservative system.
It is worth mentioning that the global existence result in the present paper is different from that obtained in [16], which is not uniform with respect to \(\varepsilon _1\) and \(\varepsilon _2\). More precisely, there are terms \(\partial _tq\) and \(\partial _t\tau \) in the definition of the energy in [16]. Because of boundary layers in the limit as \(\varepsilon \rightarrow 0\), such an energy cannot be uniformly bounded with respect to the relaxation parameters. In order to avoid this situation, in the proof of our results, the energy contains only terms of derivative of solutions with respect to x.
Theorem 1.1
(Uniform global existence). Let \(s\ge 2\) be an integer and \((\rho _0^\varepsilon -1, u_0^\varepsilon , \theta _0^\varepsilon -1, q_0^\varepsilon , \tau _0^\varepsilon )\in H^s({\mathbb {R}})\). There exist two positive constants \(\delta \) and C, independent of \(\varepsilon _1\) and \(\varepsilon _2\), such that if
then for all \(\varepsilon _1,\varepsilon _2\in (0,1]\), the Cauchy problem (1.11)-(1.12) together with (1.14)-(1.15) admits a unique global smooth solution \((\rho ^\varepsilon , u^\varepsilon ,\theta ^\varepsilon , q^\varepsilon , \tau ^\varepsilon )\) satisfying
and
where \(\Vert \cdot \Vert _k\) denotes the usual norm of \(H^k({\mathbb {R}})\).
Theorem 1.2
(Global convergence). Let \(\varepsilon =(\varepsilon _1,\varepsilon _2)\) and \((\rho ^\varepsilon , u^\varepsilon , \theta ^\varepsilon , q^\varepsilon , \tau ^\varepsilon )\) be the global solution obtained in Theorem 1.1, then there exist functions \(({\bar{\rho }},{\bar{u}},{\bar{\theta }})\in L^\infty ({\mathbb {R}}^+;H^s({\mathbb {R}}))\) and \(({\bar{q}},{\bar{\tau }})\in L^2({\mathbb {R}}^+; H^s({\mathbb {R}}))\), such that, as \(\varepsilon \rightarrow 0\) up to subsequences,
where \(({\bar{\rho }},{\bar{u}},{\bar{\theta }})\) is the solution to the one-dimensional full compressible Navier-Stokes equations (1.13) for the ideal fluid (1.2), with initial value \(({\bar{\rho }}_0,{\bar{u}}_0,{\bar{\theta }}_0)\) which is the weak limit of \((\rho _0^\varepsilon , u_0^\varepsilon , \theta _0^\varepsilon )\) up to subsequences. Moreover,
The rest of the paper is organized as follows. In the next section, we prove results on the hyperbolic structure for system (1.11). These results are crucial in the proof of the above theorems. Section 3 is devoted to uniform global estimates. The proof of the theorems are completed in the last section.
2 Symmetrizable Hyperbolicity
In what follows, \(s\ge 2\) denotes an integer. Let \(\varepsilon =(\varepsilon _1,\varepsilon _2)\) and C be a generic positive constant independent of \(\varepsilon \) and any time. We assume that \(\varepsilon _1,\varepsilon _2\in (0,1]\). For a integer \(k\ge 1\), we denote by \(\Vert \cdot \Vert _k\), \(\Vert \cdot \Vert \) and \(\Vert \cdot \Vert _\infty \) the norms of the usual Sobolev spaces \(H^k({\mathbb {R}})\), \(L^2({\mathbb {R}})\) and \(L^\infty ({\mathbb {R}})\), respectively. The inner product in \(L^2\left( {\mathbb {R}}\right) \) is denoted by \(\big \langle \cdot ,\,\cdot \big \rangle \). In the proof we will frequently use the fact that the embedding from \(H^l({\mathbb {R}})\) to \(L^\infty ({\mathbb {R}})\) is continuous for all integers \(l\ge 1\).
For simplicity, the dependence of solution on the parameters \(\varepsilon _1\) and \(\varepsilon _2\) is not expressed explicitly. We want to write (1.11) into a first-order quasilinear system of variables \((\rho ,u,\theta ,q,\tau )\). First, it is clear that by using (1.15) and the first equation in (1.11), the second equation in (1.11) is equivalent to
Similarly, by using the first two equations in (1.11), the third equation in (1.11) is equivalent to
This equation can be further treated by using equations (1.3), (1.14) and (1.15). Indeed,
and
hence,
Combining the last equation with (1.3), (2.1) and the first equation in (1.11), we have
Therefore, system (1.11) is equivalent to
Let \(D_0(\varepsilon )\) be a diagonal matrix defined by
Then system (2.2) is written as
where
and
We define
with
Then straightforward calculations give
where
It is clear that \(A_0(V)\) and \({\tilde{A}}(V)\) are symmetric. Let \(T>0\) and
We denote
In Theorem 1.1, estimate (1.16) implies that \({\mathscr {E}}_T\) is uniformly sufficiently small when \({\mathscr {E}}(0)\) is, although the \(L^\infty (0,T;H^s({\mathbb {R}}))\) norm of q and \(\tau \) may not be uniformly small with respect to \(\varepsilon _1\) and \(\varepsilon _2\). The following result shows that \(A_0(V)\) is positive definite when \({\mathscr {E}}_T\) is sufficiently small.
Lemma 2.1
Let \(W\in C([0,T];H^s({\mathbb {R}}))\) be the smooth solution to (2.2) with (1.14)-(1.15) and (1.12). Then there exist constants \(\delta >0\) and \(c_1>0\), independent of \(\varepsilon _1\) and \(\varepsilon _2\), such that if \({\mathscr {E}}_T^{1/2}\le \delta \), we have
and \(A_0(V)\) is positive definite and then system (2.2) is symmetrizable hyperbolic.
Proof
When \({\mathscr {E}}_T\) is sufficiently small, \(\rho \) and \(\theta \) are sufficiently close to 1, then they have uniform positive upper and lower bounds. From the state equations (1.14) and (1.15), we have
which imply (2.5) since \(\Vert \varepsilon _1 q\Vert _\infty \le {\mathscr {E}}_T\) which is sufficiently small. Besides,
which imply (2.6).
Moreover,
Then (2.5) implies that the above determinate is positive for sufficiently small \({\mathscr {E}}_T\). From (2.6) and \(\varepsilon _2\le 1\), we also have \(\Vert p_\tau \Vert _s\le C\delta \). Therefore, \(A_0(V)\) is positive definite from its explicit expression. \(\square \)
Applying the theory on the symmetrizable hyperbolic system, Lemma 2.1 implies the local existence of smooth solutions to the Cauchy problem (2.2) with (1.14)-(1.15) and (1.12) from the classical iteration technique and fixed point theorems, see for instance [18, 21, 24].
3 Uniform Global Estimates
Let \(T>0\) and \(W=(N,u,\Theta ,\varepsilon _1q,\varepsilon _2\tau )^\top \) be the unique local smooth solution to (2.2) with (1.14)-(1.15) and (1.12), defined on [0, T]. We assume that \({\mathscr {E}}_T\) defined in (2.4) is sufficiently small. This gives rise to the rational assumption that
We want to establish the uniform global estimate for W with respect to the parameters \(\varepsilon _1\) and \(\varepsilon _2\) and the time T. For this purpose, we introduce the following dissipative energy
3.1 \(L^2\)-estimate
The \(L^2\) estimate relies on the existence of a strictly convex entropy and its corresponding entropy flux.
Lemma 3.1
For system (2.2), there exists a strictly convex entropy \(\eta \) and its corresponding entropy flux \(\Psi \) satisfying
where
and
In addition, if \({\mathscr {E}}_T\) is sufficiently small, the following \(L^2\)-energy estimate holds
Proof
The entropy-entropy flux identity (3.2) was established in [16]. We now prove (3.3). By the Taylor’s expansion at \((\rho ,\theta )=(1,1)\) and (3.1), we have
where \({\hat{\rho }}\) is between \(\rho \) and 1, and \({\hat{\theta }}\) is between \(\theta \) and 1. By Lemma 2.1, this implies that there exists a constant \(c_2>0\), such that
where \(|\cdot |\) is the Euclidean norm in \({\mathbb {R}}^5\). Integrating (3.2) over [0, t] for \(t\in [0,T]\) implies (3.3). \(\square \)
3.2 Higher Order Estimates
Let l be an integer with \(1\le l\le s\). Applying \(\partial _x^l\) to both sides of the equation (2.3) yields
where
Then,
Taking the inner product of the above system with \(\partial _x^l V\) in \(L^2({\mathbb {R}})\), we have
We deal with (3.4) term by term in a series of lemmas as follows.
Lemma 3.2
There exists positive constants \(c_3\) and \(c_4\) such that
Proof
By the definition of \({\tilde{A}}_0(V)\) and B together with (2.8), we have
which implies that
Obviously,
For the second term, noticing that \(\dfrac{\rho \theta p_\theta }{\kappa }\) has a uniform lower bound when \({\mathscr {E}}_T\) is sufficiently small. Besides, the last estimate in (2.6) implies that \(\dfrac{2\rho qp_q}{\kappa }\) is small and consequently, there exists a positive constant \(c_3\) such that
Similarly, (3.1) implies that \(1-p_\tau \ge \dfrac{1}{2}\). Therefore, there exists a positive constant \(c_4\) such that
Combining all these estimates yields (3.5). \(\square \)
Lemma 3.3
It holds
Proof
We denote
and
where
Obviously,
From (1.14) and (1.15), we see that \(A^{11}_0(V)\) is a smooth function of \((N,\Theta ,\varepsilon ^2_1q^2)\). Moreover, from (2.2) together with the Cauchy-Schwarz inequality, we have
Hence,
Moreover,
Since \({\mathscr {E}}_T\) is uniformly small, we also have
Therefore,
Next, a straightforward calculation yields
From (2.7), we have
Then (2.2) together with the bound of \({\mathscr {E}}_T\) yields
Therefore,
Finally,
Similarly to the estimates above, we have
and
which imply that
This proves the lemma. \(\square \)
Lemma 3.4
It holds
Proof
From (1.14) and (1.15) and the explicit expression of \({\tilde{A}}(V)\), we see that all elements of \({\tilde{A}}(V)\) are smooth functions of \((\rho ,u,\theta ,\varepsilon ^2_1q^2,\varepsilon _2^2\tau ^2)\) except \(J_1\). Noticing that \(J_1\) appears at the position of the third line and the third column of \({\tilde{A}}(V)\) and
we obtain
This proves (3.7). \(\square \)
Lemma 3.5
It holds
Proof
By the definition of \({\tilde{A}}_0(V)\) and F, we have
It follows that
which proves (3.8). \(\square \)
Lemma 3.6
It holds
Proof
Recall that
Similarly to the proof of Lemma 3.4, it is easy to see that all elements of A(V) are smooth functions of \((\rho ,u,\theta ,\varepsilon ^2_1q^2,\varepsilon _2^2\tau ^2)\) except the element at the position of the third line and the third column. We denote by \(J_3\) this element. Then
and
The only two terms that contain \(J_3\) are the following
which can be treated similarly to the proof of Lemma 3.4 and are obviously bounded by \(C{\mathscr {E}}_T^{1/2}{\mathscr {D}}(t)\).
On the other hand, each element of \({\tilde{A}}_0(V)\) is uniformly bounded in \(L^\infty ([0,T]\times {\mathbb {R}})\) except for the element at the position of the third line and the fourth column. This element is \(-\frac{1}{\kappa }(\rho ^2q e_\theta )\) which only touches the fourth line of \(K_l\) in the product \({\tilde{A}}_0(V)K_l\). Now the only nonzero element on the fourth line of \(K_l\) is \(\varepsilon ^2_1(u\partial _x^{l+1}q-\partial _x^l(u\partial _xq))\). Then the fourth component of \({\tilde{A}}_0(V)K_l\) is \(-\frac{\varepsilon ^2_1}{\kappa }(\rho ^2q e_\theta )(u\partial _x^{l+1}q-\partial _x^l(u\partial _xq))\). By the Moser-type calculus inequalities (see [24] for instance), we have
The estimates for the other terms can be easily obtained. This proves (3.9). \(\square \)
3.3 Dissipative Estimates for \(\partial _xN\), \(\partial _xu\) and \(\partial _x\Theta \)
Lemma 3.7
(Dissipative estimates for \(\partial _x\Theta \)). It holds
where \(\nu >0\) is a small positive constant to be determined later.
Proof
Let m be an integer with \(0\le m\le s-1\). Applying \(\partial _x^m\) to the fourth equation in (2.2) and making the inner product of the resulting equation with \(\partial _x^{m+1}\Theta \) in \(L^2({\mathbb {R}})\), we have
Obviously,
Therefore,
By using the third equation in (2.2) and an integration by parts, we have
Obviously,
and
Combining these estimates, we arrive at
This together with (3.11) yields
Adding this inequality for all \(0\le m\le s-1\) yields (3.10). \(\square \)
Lemma 3.8
(Dissipative estimates for \(\partial _xu\)). It holds
where \(\nu >0\) is a small positive constant to be determined later.
Proof
For \(0\le m\le s-1\), applying \(\partial _x^m\) to the fifth equation in (2.2) and making the inner product of the resulting equation with \(\partial _x^{m+1}u\) in \(L^2({\mathbb {R}})\), we have
Obviously,
Therefore,
By using the second equation in (2.2) and an integration by parts, we have
Substituting this estimate into (3.13) and adding the resulting equation for all \(0\le m\le s-1\) yield (3.12). \(\square \)
Lemma 3.9
(Dissipative estimates for \(\partial _xN)\). It holds
Proof
We first write the second equation in (2.2) as
For \(m\le s-1\), applying \(\partial _x^m \) to the above equation and making the inner product of the resulting equation with \(\partial _x^{m+1}N\) in \(L^2({\mathbb {R}})\), we have
Noticing (3.1), we have
It is clear that
and by the Moser-type calculus inequalities,
Moreover, by the mass equation in (2.2) and the integration by parts, we obtain
Combining all these estimates, we arrive at
Adding the above estimate for all \(m\le s-1\) yields (3.14). \(\square \)
4 Proof of Theorems 1.1-1.2
Lemma 4.1
(Final energy estimate). If \({\mathscr {E}}_T\) is sufficiently small, then
Proof
Combining (3.4) and Lemmas 3.2-3.6, and adding for all \(1\le l\le s\), we arrive at
Since \(A_0(V)\) is positive definite with respect to W, there exists a constant \(c_5>0\), independent of \(\varepsilon _1\) and \(\varepsilon _2\), such that
Integrating (4.2) over [0, T] together with (3.3) yields
On the other hand, combining the estimates in Lemmas 3.7-3.9, we have
where \(\alpha >0\) is a constant. We choose \(\nu >0\) and \(\alpha >0\) sufficiently small such that
It follows that there is a constant \(a_0>0\), independent of \(\varepsilon _1\) and \(\varepsilon _2\), such that
It is clear that
Integrating (4.4) over [0, T] yields
This inequality together with (4.3) yields
which implies (4.1) since \({\mathscr {E}}_T\) is sufficiently small. \(\square \)
Proof of Theorem 1.1
The estimate in Lemma 4.1 shows that the smooth solution W is uniformly bounded in \(L^\infty ([0,T];H^s({\mathbb {R}}))\) with respect to \(\varepsilon \) and T. By the bootstrap principle, it yields uniformly global solution. In particular, this estimate gives (1.16). \(\square \)
Proof of Theorem 1.2
From (1.11), \((\rho ^\varepsilon , u^\varepsilon , \theta ^\varepsilon , q^\varepsilon , \tau ^\varepsilon )\) satisfies the following system
in \({\mathbb {R}}^+\times {\mathbb {R}}\), where
The uniform estimate (1.16) implies that the sequences \((\rho ^\varepsilon -1)_{\varepsilon }\), \((u^\varepsilon )_{\varepsilon }\) and \((\theta ^\varepsilon -1)_{\varepsilon }\) are bounded in \(L^\infty ({\mathbb {R}}^+;H^s({\mathbb {R}}))\) and the sequence \((q^\varepsilon )_{\varepsilon }\) and \((\tau ^\varepsilon )_{\varepsilon }\) are bounded in \(L^2({\mathbb {R}}^+;H^s({\mathbb {R}}))\). It follows that there exist functions \(({\bar{\rho }},{\bar{u}},{\bar{\theta }})\in L^\infty ({\mathbb {R}}^+;H^s({\mathbb {R}}))\) and \(({\bar{q}},{\bar{\tau }})\in L^2({\mathbb {R}}^+; H^s({\mathbb {R}}))\), such that (1.17)-(1.18) hold. In addition, as \(\varepsilon \rightarrow 0\),
and
Moreover, from the first three equations in (4.5), it is easy to see that \((\partial _t\rho ^\varepsilon )_{\varepsilon }\), \((\partial _tu^\varepsilon )_{\varepsilon }\) and \((\partial _t\theta ^\varepsilon )_{\varepsilon }\) are bounded in \(L^2({\mathbb {R}}^+;H^{s-1}({\mathbb {R}}))\). Hence, by a classical compactness theorem [40], for all \(T>0\), \((\rho ^\varepsilon )_{\varepsilon },(u^\varepsilon )_{\varepsilon }\) and \((\theta ^\varepsilon )_{\varepsilon }\) are relatively compact in \(C([0,T];H_{loc}^{s-1}({\mathbb {R}}))\). As a consequence, as \(\varepsilon \rightarrow 0\), up to subsequences,
This is sufficient to pass the limit in (4.5)-(4.6) in the sense of distributions and to obtain the Navier-Stokes equations for the ideal fluid. This ends the proof of Theorem 1.2. \(\square \)
References
Boillat, G.: Sur l’existence et la recherche d’équations de conservation supplémentaires pour les systèmes hyperboliques. C. R. Acad. Sci. Paris Sér. A 278, 909–912 (1974)
Bresch, D., Prange, C.: Newtonian limit for weakly viscoelastic fluid flows. SIAM J. Math. Anal. 46(2), 1116–1159 (2014)
Cattaneo, C.: Sulla conduzione del calore. Atti Sem. Mat. Fis. Univ. Modena 3, 83–101 (1949)
Cattaneo, C.: Sur une forme de l’équation de la chaleur éliminant le paradoxe d’une propagation instantanée. C. R. Acad. Sci. Paris 247, 431–433 (1958)
Chen, P.J., Gurtin, M.E.: On second sound in materials with memory. Z. Angew. Math. Phys. 21(2), 232–241 (1970)
Christov, C.I.: On frame indifferent formulation of the Maxwell–Cattaneo model of finite-speed heat conduction. Mech. Res. Commun. 36(4), 481–486 (2009)
Christov, C.I., Jordan, P.M.: Heat conduction paradox involving second-sound propagation in moving media. Phys. Rev. Lett. 94(15), 154301 (2005)
Coleman, B.D., Fabrizio, M., Owen, D.R.: On the thermodynamics of second sound in dielectric crystals. Arch. Ration. Mech. Anal. 80(2), 135–158 (1982)
Coleman, B.D., Hrusa, W.J., Owen, D.R.: Stability of equilibrium for a nonlinear hyperbolic system describing heat propagation by second sound in solids. Arch. Ration. Mech. Anal. 94(3), 267–289 (1986)
Coulaud, O., Hachicha, I., Raugel, G.: Hyperbolic quasilinear Navier–Stokes equations in \({\mathbb{R}}^2\). J. Dynam. Diff. Equa. (2021)
Feireisl, E.: Dynamics of viscous compressible fluids. Oxford Lecture Series in Mathematics and its Applications, vol. 26. Oxford University Press, Oxford (2004)
Godunov, S.K.: An interesting class of quasi-linear systems (Russian). Dokl. Akad. Nauk SSSR 139, 521–523 (1961)
Hoff, D.: Global existence for 1d, compressible, isentropic Navier–Stokes equations with large initial data. Trans. Am. Math. Soc. 303(1), 169–181 (1987)
Hu, Y., Racke, R.: Compressible Navier–Stokes equations with hyperbolic heat conduction. J. Hyper. Diff. Equa. 13(2), 233–247 (2016)
Hu, Y., Racke, R.: Compressible Navier–Stokes equations with revised Maxwell’s law. J. Math. Fluid Mech. 19(1), 77–90 (2017)
Hu, Y., Racke, R.: Hyperbolic compressible Navier–Stokes equations. J. Diff. Equa. 269(4), 3196–3220 (2020)
Kanel, Ja. I.: A model system of equations for the one-dimensional motion of a gas. Diff. Equa. (in Russian) 4, 721–734 (1968)
Kato, T.: The Cauchy problem for quasi-linear symmetric hyperbolic systems. Arch. Ration. Mech. Anal. 58(3), 181–205 (1975)
Kawashima, S.: Systems of a hyperbolic–parabolic composite type, with applications to the equations of magnetohydrodynamics. Ph.D thesis, Kyoto Uninversity (1983)
Kazhikhov, A.V., Shelukhin, V.V.: Unique global solution with respect to time of initial-boundary value problems for one-dimensional equations of a viscous gas. J. Appl. Math. Mech. 41(2), 273–282 (1977)
Lax, P.D.: Hyperbolic systems of conservation laws and the mathematical theory of shock waves. volume 11. SIAM Regional Conf. Lecture, Philadelphia (1973)
Li, Y., Peng, Y.J., Zhao, L.: Convergence rate from hyperbolic systems of balance laws to parabolic systems. Appl. Anal. 100(5), 1079–1095 (2021)
Lions, P.-L.: Mathematical Topics in Fluid Mechanics: Volume 2: Compressible Models, vol. 2. Clarendon Press, Oxford (1998)
Majda, A.: Compressible Fluid Flow and Systems of Conservation Laws in Several Space Variables, vol. 53. Springer, New York (1984)
Majda, A., Bertozzi, A.: Vorticity and Incompressible Flow. Cambridge Texts in Applied Mathematics, vol. 27. Cambridge University Press, Cambridge (2002)
Masmoudi, N.: Examples of singular limits in hydrodynamics. In: Handbook of Differential Equations: Evolutionary Equations, pages 195–275. Elsevier (2007)
Matsumura, A., Nishida, T.: The initial value problem for the equations of motion of viscous and heat-conductive gases. J. Math. Kyoto Univ. 20(1), 67–104 (1980)
Maxwell, J.C.: IV. on the dynamical theory of gases. Philos. Trans. R. Soc. London 157, 49–88 (1867)
Molinet, L., Talhouk, R.: Newtonian limit for weakly viscoelastic fluid flows of Oldroyd type. SIAM J. Math. Anal. 39(5), 1577–1594 (2008)
Peng, Y.J.: Relaxed Euler systems and convergence to Navier–Stokes equations. Ann. Inst. H. Poincaré Anal. Non Linéaire 38(2), 369–401 (2021)
Peng, Y.J., Wasiolek, V.: Parabolic limit with differential constraints of first-order quasilinear hyperbolic systems. Ann. Inst. H. Poincaré Anal. Non Linéaire 33(4), 1103–1130 (2016)
Peng, Y.J., Wasiolek, V.: Uniform global existence and parabolic limit for partially dissipative hyperbolic systems. J. Diff. Equa. 260(9), 7059–7092 (2016)
Racke, R., Saal, J.: Hyperbolic Navier–Stokes equations I: Local well-posedness. Evol. Equ. Control Theory 1 1(1), 195–215 (2012)
Racke, R., Saal, J.: Hyperbolic Navier–Stokes equations II: Global existence of small solutions. Evol. Equ. Control Theory 1 1(1), 217–234 (2012)
Renardy, M., Hrusa, W.J., Nohel, J.A.: Mathematical problems in viscoelasticity, volume 35 of Pitman Monographs and Surveys in Pure and Applied Mathematics. Longman Scientific & Technical, Harlow; Wiley, New York (1987)
Saut, J.-C.: Some remarks on the limit of viscoelastic fluids as the relaxation time tends to zero. In: Kröner, E., Kirchgässner, K. (eds.) Trends in Applications of Pure Mathematics to Mechanics, pp. 364–369. Springer, Berlin (1986)
Schöwe, A.: A quasilinear delayed hyperbolic Navier–Stokes system: global solution, asymptotics and relaxation limit. Methods Appl. Anal. 19(2), 99–118 (2012)
Schöwe, A.: Blow-up results to certain hyperbolic model problems in fluid mechanics. Nonlinear Anal. 144, 32–40 (2016)
Shizuta, Y., Kawashima, S.: Systems of equations of hyperbolic-parabolic type with applications to the discrete Boltzmann equation. Hokkaido Math. J. 14(2), 249–275 (1985)
Simon, J.: Compact sets in the space \(L^p(0, T;B)\). Ann. Mat. Pura Appl. (4) 146, 65–96 (1987)
Tarabek, M.A.: On the existence of smooth solutions in one-dimensional nonlinear thermoelasticity with second sound. Q. Appl. Math. 50(4), 727–742 (1992)
Yong, W.-A.: Newtonian limit of Maxwell fluid flows. Arch. Ration. Mech. Anal. 214(3), 913–922 (2014)
Zhao, L., Xi, S.: Convergence rate from systems of balance laws to isotropic parabolic systems, a periodic case. Asymptot. Anal. 124, 163–198 (2021)
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
There is no conflict of interest.
Additional information
Communicated by G. P. Galdi.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Peng, YJ., Zhao, L. Global Convergence to Compressible Full Navier–Stokes Equations by Approximation with Oldroyd-Type Constitutive Laws. J. Math. Fluid Mech. 24, 29 (2022). https://doi.org/10.1007/s00021-022-00669-4
Accepted:
Published:
DOI: https://doi.org/10.1007/s00021-022-00669-4
Keywords
- Full Navier–Stokes equations
- Non-Newtonian fluid
- Relaxed Euler systems
- Oldroyd derivative
- Global convergence