Abstract
We prove the energy equality of MHD system in the space founded by Cheskidov et al. (Nonlinearity 21:1233–1252, 2008) and Berselli and Chiodaroli (Nonlinear Anal 192:111704, 2020). It is clarified that the energy equality is established for a larger class of the magnetic field than that of velocity field. Most of the cases, we deal with the energy equality of MHD system in bounded domain. On the other hand, if the spacial integrability exponents of the weak solution are large, it is necessary to use the Besov space, which is suitable for us to handle freely derivatives of the nonlinear convection term. Only in this case we deal with the energy equality of MHD system in the whole space. Our result covers most of previous theorems on validity of the energy equality on the Navier–Stokes equations.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
In this paper, we consider the following MHD system in a bounded domain \(\Omega \subset {\mathbb {R}}^3\) with the smooth boundary \(\partial \Omega \):
where \(u=u(x,t) = (u_1(x, t), u_2(x, t), u_3(x, t)),\) \(B=B(x,t) = (B_1(x, t), B_2(x, t), B_3(x, t))\) and \(p = p(x, t)\) denote the unknown velocity, magnetic field and the pressure, respectively, while \(u_0=u_0(x) = (u_{0, 1}(x), u_{0,2}(x), u_{0,3}(x))\) and \(B_0=B_0(x) = (B_{0, 1}(x), B_{0,2}(x), B_{0,3}(x))\) denote the given initial velocity and magnetic field, respectively. Notice that \(\nu \) is the unit outer normal to \(\partial \Omega .\) For smooth solutions u and B of (1.1), the following energy equality holds:
Our aim is to show the energy equality for a larger class of solutions.
The energy conservation law of MHD system has been developed together with that of the Navier–Stokes equations since both equations have a similar non-linear structure. Let us first recall previous results on the Navier–Stokes equations. The pioneer paper work by Shinbrot [11] showed that the energy equality holds for the solution in \(L^p(0,T; L^q(\Omega ))\) for \(2/p+2/q \leqq 1\) with \(4 \leqq q.\) It should be noted that such a class given by Shinbrot is larger than that of the scaling invariant space given by Serrin [10]. Recently, Cheskidov–Constantin–Friedlander–Shvydkoy [3] obtained an elegant result in \({\mathbb {R}}^3\) in the largest class \(L^3(0,T; B^{1/3}_{3,\infty }({\mathbb {R}}^3)).\) From this result and the interpolations with \(L^\infty (0,T; L^2({\mathbb {R}}^3)) \cap L^2(0,T;H^1({\mathbb {R}}^3)),\) the energy equality holds for the solution in the following class:
where \(B^\alpha _{q,\infty }{({\mathbb {R}}^3)}\) denotes the inhomogeneous Besov space in \({\mathbb {R}}^3.\) The criteria (1.3)–(1.5) and the interpolation method are mentioned in Cheskidov–Luo [4, Lemma 2.1]). Berselli–Chiodaroli [1] showed the energy equality under the following conditions which are included in (1.3)–(1.5) for exponents p, q but deal with a bounded domain:
As for another extension of these results, to enlarge the space in [3], Cheskidov and Luo [4] introduced the weak \(L^p\)-space including \(0<p<1\) as follows:
where \(L^p_w\) denote the weak \(L^p\) space on (0, T). Let us introduce the class
for \(2/p+3/q-1-\alpha =\varepsilon (p,q,\alpha ).\) In the case \(\varepsilon (p,q,\alpha )=0,\) the class (1.11) coincides with the scaling invariant space. From viewpoint of the scaling invariant space, it is easy to prove (1.2) in the case \(\varepsilon (p,q,\alpha ) \leqq 0.\) Hence, the essential problem for the validity of the energy equality is to take \(\varepsilon (p,q,\alpha )>0.\)
Concerning ideal MHD system, the result corresponding to Constantin–E–Titi [5] in the Euler equation was handled by Caflisch–Klapper–Steele [2]. According to the paper, the energy equality holds if \(u \in C([0,T];B^{\alpha _1}_{3,\infty }({\mathbb {T}}^3))\) and \(B \in C([0,T];B^{\alpha _2}_{3,\infty }({\mathbb {T}}^3)),\) where the Hölder exponents \(\alpha _1\) and \(\alpha _2\) satisfy \(\alpha _1>1/3\) and \(\alpha _1 + 2 \alpha _2 > 1.\) In the whole space \({\mathbb {R}}^3,\) the same result on the critical case of Caflisch et al. [2] was obtained by Kang–Lee [8]. Namely, the energy equality for the ideal MHD system holds if \(u \in L^3([0,T];B^{\alpha _1}_{3,c({\mathbb {N}})}({\mathbb {R}}^3))\) and \(B \in L^3([0,T];B^{\alpha _2}_{3,c({\mathbb {N}})}({\mathbb {R}}^3)),\) where the Hölder exponents \(\alpha _1\) and \(\alpha _2\) satisfy \(\alpha _1 \geqq 1/3\) and \(\alpha _1 + 2 \alpha _2 \geqq 1.\) Recently, in the case of the bounded domain \(\Omega ,\) the result corresponding to Caflisch et al. [2] was obtained by Wang–Zuo [15].
Concerning MHD system with viscosity and magnetic diffusion, He–Xin [6] showed the validity of the energy equality for u in the Serrin class even though B is in the class \(L^\infty (0, T; L^2({\mathbb {R}}^3)) \cap L^2(0, T; H^1({\mathbb {R}}^3))\) given by Leray [9] and Hopf [7]. The same result on the critical case \(u \in L^\infty (0,T; L^3({\mathbb {R}}^3))\) of Serrin’s was obtained by Yong–Jiu [16]. On the other hand, the critical case \(L^2_w(0,T;BMO_{loc}({\mathbb {R}}^3))\) of Serrin’s was handled by Tan–Wu [12]. They introduce the function space \(L^2_w(0,T;BMO_{loc}({\mathbb {R}}^3))\) which is bigger than \(L^2(0,T;L^\infty ({\mathbb {R}}^3))\) but they assume that B must also belong to \(L^2_w(0,T;BMO_{loc}({\mathbb {R}}^3)).\) Recently, Wang–Zuo [14] dealt with the corresponding result to He–Xin [6] in the bounded domain \(\Omega .\) They also dealt with the critical case of Shinbrot [11]. The result corresponding to Shinbrot [11] and Taniuchi [13] was also handled by Zeng [17]. However, in the Shinbrot–Taniuchi condition dealt in Wang–Zuo [14] and Zeng [17], they assumed that B belongs to the same class of u. Since the nonlinear structures of u and B in the MHD system are the same, it is expected that if B belongs to the same class of u satisfying the energy equality of the Navier–Stokes equations, then we have (1.2). Hence, it seems to be an interesting question to find a larger class of B than that of u so that we may obtain (1.2).
In this paper, we relax the condition on B in such a way that the energy equality holds for B in a larger class than that of Berselli–Chiodaroli [1] for the bounded domain \(\Omega \) and Cheskidov–Constantin–Friedlander–Shvydkoy [3] for the whole space \({\mathbb {R}}^3.\)
This paper is organized as follows. In Sect. 2, we state our main results. Sections 3 and 4 are devoted to the proof of our main theorems.
2 Main results
Theorem 2.1
Let u and B be weak solutions of (1.1) in \(L^\infty (0,T;L^2(\Omega )) \cap L^2(0,T;H^1(\Omega )).\)
-
(i)
Case \(3/2 < q \leqq 9/5{:}\) If u and B satisfy
$$\begin{aligned}{} & {} {\textrm{rot}}\ u \in L^p(0,T; L^q(\Omega ))\quad \text {for}\ \frac{2}{p}+\frac{3}{q}-2=2-\frac{3}{q},\end{aligned}$$(2.1)$$\begin{aligned}{} & {} \begin{array}{c} {\textrm{rot}}\ B \in L^\sigma (0,T;L^\tau (\Omega ))\quad \text {for}\ \dfrac{2}{\sigma }+\dfrac{3}{\tau }-2=2-\dfrac{3}{\tau },\\ \quad \text {with}\ \dfrac{3}{2} < \tau \leqq \dfrac{9}{5}+\varepsilon _1(q), \end{array} \end{aligned}$$(2.2)then the energy equality (1.2) holds, where
$$\begin{aligned} 0\leqq \varepsilon _1(q)=\frac{6q}{5q-3}-\frac{9}{5}. \end{aligned}$$Notice that \(\varepsilon _1(9/5) = 0\) and \(0<\varepsilon _1(q)\) for \(3/2< q < 9/5.\)
-
(ii)
Case \(9/5 \leqq q < 3{:}\) If u and B satisfy
$$\begin{aligned}{} & {} {\textrm{rot}}\ u \in L^p(0,T; L^q(\Omega ))\quad \text {for}\ \frac{2}{p}+\frac{3}{q}-2=\frac{3}{5q},\end{aligned}$$(2.3)$$\begin{aligned}{} & {} \begin{array}{c} {\textrm{rot}}\ B \in L^\sigma (0,T;L^\tau (\Omega ))\quad \text {for}\ \dfrac{2}{\sigma }+\dfrac{3}{\tau }-2=\dfrac{3}{5\tau },\\ \quad \text {with}\ \dfrac{9}{5}-\varepsilon _2(q) \leqq \tau < 3, \end{array} \end{aligned}$$(2.4)then the energy equality (1.2) holds, where
$$\begin{aligned} 0 \leqq \varepsilon _2(q)=\frac{9}{5}-\frac{6q}{5q-3}. \end{aligned}$$
Notice that \(\varepsilon _2(9/5) = 0\) and \(0<\varepsilon _2(q)\) for \(9/5< q < 3.\)
Since \(q=3\) exhibits a threshold exponent for validity of (1.2), it is suitable to introduce the Besov space to obtain a new criteria.
Theorem 2.2
Let u and B be weak solutions of (1.1) in \(L^\infty (0,T;L^2({\mathbb {R}}^3)) \cap L^2(0,T;H^1({\mathbb {R}}^3)).\) Assume that \(3 \leqq q <\infty .\) If u and B satisfy
then the energy equality (1.2) holds, where \(0 < \varepsilon _2(q).\)
Remark 1
1. The classes (2.1), (2.3) and (2.5) of u are same as those in recent works [1, 3] for the Navier–Stokes equations. However, our classes (2.2), (2.4) and (2.6) of B are larger than those of them. Especially, if we take u in the scaling invariant spaces \(L^1(0,T;W^{1,\infty })\) and \(L^\infty (0,T; W^{1,3/2}),\) then B reaches the Leray-Hopf class. Hence, we may regard our Theorems 2.1 and 2.2 as a generalization of He–Xin [6].
2. The classes (2.1) and (2.2) correspond to (1.6). The classes (2.3) and (2.4) also correspond to (1.7). Notice that the classes (2.5) and (2.6) dose not relate to (1.8), but corresponds to (1.4) in the case \(3<q.\) More precisely, (1.8) represents the nonlinear relation between quantities 2/p and 3/q by means to the inhomogeneous term \(4/(q+2)\) on the RHS. On the other hand, our advantage of (2.5) and (2.6) exhibit the linear relation between 2/p and 3/q, which enables us to obtain (1.2) in the Leray-Hopf class for B when u belongs to the Serrin class. In (2.5) and (2.6), we should also emphasize the advantage to make use of the Besov space, which is suitable for us to handle freely derivatives of the nonlinear convection term.
3. The class in our Theorems 2.1 and 2.2 are not embedded in the Kang–Lee’s class which deal with the ideal MHD in \({\mathbb {R}}^3\) by the interpolations with the Leray–Hopf class. Indeed, when u belongs to the class near the critical cases \(L^1(0,T;B^1_{\infty ,\infty }({\mathbb {R}}^3))\) or \(L^\infty (0,T; W^{1,3/2}({\mathbb {R}}^3)),\) our results are not embedded in the class of Kang–Lee [8]. For instance, we consider the critical case \(u \in L^1(0,T;B^1_{\infty ,\infty }({\mathbb {R}}^3))\) and \(B \in L^\infty (0,T; B^1_{6/5, \infty }({\mathbb {R}}^3))\) in Theorem 2.2. By the interpolation method mentioned in Cheskidov–Luo [4], we see that \(\alpha _2=1/6\) is the largest exponent \(\alpha _2\) in which the following embedding holds:
Then, we choose the smallest exponent \(\alpha _1 = 2/3\) satisfying \(1 \leqq \alpha _1 + 2 \alpha _2.\) Now, we consider the following imbedding:
with the conditions
However, there exist no exponents s, p, \(\theta _1\) and \(\theta _2\) satisfying the imbedding and conditions above. Indeed, these conditions imply \(1/3 \geqq \theta _1\) and \(\theta _1 \geqq 1,\) which is a contradiction. Hence, our class in Theorem 2.2 is not embedded in the Kang–Lee’s class by the interpolations with the Leray–Hopf class. In the case of \(u \in L^\infty (0,T; W^{1,3/2}({\mathbb {R}}^3))\) and \(B \in L^2(0,T; W^{1,2}({\mathbb {R}}^3))\) in Theorem 2.1, our class is also not embedded in the class in Kang–Lee [8] by the interpolations with the Leray–Hopf class. Moreover, our conditions can deal with not only the whole space \({\mathbb {R}}^3\) but also the bounded domain \(\Omega .\)
3 Preliminaries
3.1 Notations
In this paper, \(p'\) and \(p^*\) denote the Hölder and Sobolev exponents defined by
We also use the inner product \((\cdot ,\cdot )\) and norms \(\Vert f\Vert _{L^p L^q}\) defined by
Next, let us recall the Littlewood–Paley decomposition and Besov spaces, for which we refer to [3] or [4]. Let \(\chi \in C^\infty _0(B_1(0))\) be a symmetric cut-off function satisfying \(\chi =1\) in \(B_{\frac{3}{4}}(0)\) and \(\chi =0\) in \(B_1(0)^c.\) We define \( \lambda _n=2^n, \ \phi (\xi ) = \chi (\lambda _1^{-1}\xi )-\chi (\xi )\) and \(\phi _n(\xi )=\phi (\lambda _n^{-1}\xi ).\) Then the inhomogeneous Besov space is denoted by \(B^s_{p,q}\) with the norm \(\Vert u\Vert _{B^s_{p,q}}=\Vert \lambda ^s_n\Vert \Delta _n u\Vert _p\Vert _{l^q},\) where \(\Delta _n u=\check{\phi _n}*u\ \text {for}\ 0 \leqq n\ \text {and}\ \Delta _{-1}u={\check{\chi }}*u.\) We also define \(B^s_{p,c({\mathbb {N}})}\) as
The function space \(B^s_{p,c({\mathbb {N}})}\) is introduced in Cheskidov–Constantin–Friedlander–Shvydkoy [3] and is also written as \(B^s_{p,\infty -}.\) Here, we note that for any \(1 \leqq q < \infty \) following imbedding holds;
We use notations \(u_{\leqq N},\ {{\bar{h}}},\ {{\bar{h}}}_N\) and \(K_{(\alpha )}\) defined as follows:
3.2 Estimates for non-linear terms
The proofs of our theorems are basically derived from the standard approximation argument for the energy equality as [1, 3]. Roughly speaking, since the estimates of linear terms are established in the Leray-Hopf class \(L^\infty (0,T;L^2) \cap L^2(0,T;H^1),\) the estimates of non-linear terms in following Lemmas 3.1 and 3.2 play an essential role for the proof of our theorems.
Lemma 3.1
Let \(f,\ g_1,\ g_2 \in L^\infty (0,T;L^2(\Omega )) \cap L^2(0,T;H^1(\Omega ))\) with \(f \vert _{\partial \Omega }=g_1\cdot \nu \vert _{\partial \Omega }=g_2\cdot \nu \vert _{\partial \Omega }=0.\)
-
(i)
Case \(3/2 < q \leqq 9/5{:}\) Suppose that f, \(g_1\) and \(g_2\) satisfy
$$\begin{aligned} \nabla f \in L^p(0,T;L^q(\Omega ))\quad \text {for}\ \frac{2}{p}+\frac{3}{q}-2=2-\frac{3}{q}, \\ \begin{array}{c} \nabla g_1,\ \nabla g_2 \in L^\sigma (0,T;L^\tau (\Omega ))\quad \text {for}\ \dfrac{2}{\sigma }+\dfrac{3}{\tau }-2=2-\dfrac{3}{\tau },\\ \text {with}\ \dfrac{3}{2} < \tau \leqq \dfrac{9}{5}+\varepsilon _1(q). \end{array} \end{aligned}$$ -
(i-i)
We take \(\alpha \) and \(\beta \) so that
$$\begin{aligned} \frac{1}{6}< \frac{1}{\alpha } \leqq \frac{1}{q^*},\ \frac{1}{6} < \frac{1}{\beta } \leqq \frac{1}{\tau ^*} \text { and } \frac{1}{\tau '} = \frac{1}{\alpha } + \frac{1}{\beta }. \end{aligned}$$For such \(\alpha \) and \(\beta ,\) we take \(0 < \theta ,\ \tilde{\theta }\leqq 1\) so that
$$\begin{aligned} \frac{1}{\alpha } = \frac{\theta }{q^*} + \frac{1-\theta }{6},\ \frac{1}{\beta } = \frac{{{\tilde{\theta }}}}{\tau ^*} + \frac{1-\tilde{\theta }}{6}. \end{aligned}$$Then we have the estimate
$$\begin{aligned} \left| \int ^t_0 ((f \cdot \nabla )g_1,g_2) ds \right| \leqq C \Vert \nabla f\Vert ^{\theta }_{L^pL^q} \Vert \nabla f\Vert ^{1-\theta }_{L^2L^2} \Vert \nabla g_1\Vert _{L^\sigma L^\tau } \Vert \nabla g_2\Vert ^{{{\tilde{\theta }}}}_{L^\sigma L^\tau } \Vert \nabla g_2\Vert ^{1-\tilde{\theta }}_{L^2 L^2}, \end{aligned}$$where \(C = C(q,\tau ,\alpha ,\beta ).\)
-
(i-ii)
We also take \(0 < {{\tilde{\theta }}} \leqq 1\) so that
$$\begin{aligned} \frac{1}{2q'}=\frac{{{\tilde{\theta }}}}{\tau ^*} + \frac{1-\tilde{\theta }}{6}. \end{aligned}$$Then we have the estimate
$$\begin{aligned} \left| \int ^t_0 ((g_1 \cdot \nabla ) f,g_2)ds \right| \leqq C \Vert \nabla g_1\Vert ^{{{\tilde{\theta }}}}_{L^\sigma L^\tau } \Vert \nabla g_1\Vert ^{1 - {{\tilde{\theta }}}}_{L^2 L^2} \Vert \nabla f\Vert _{L^p L^q} \Vert \nabla g_2\Vert ^{{{\tilde{\theta }}}}_{L^\sigma L^\tau } \Vert \nabla g_2\Vert ^{1 - \tilde{\theta }}_{L^2 L^2}, \end{aligned}$$where \(C = C(q,\tau ).\)
-
(ii)
Case \(9/5 \leqq q < 3\): Suppose that f, \(g_1\) and \(g_2\) satisfy
$$\begin{aligned}{} & {} \nabla f \in L^p(0,T;L^q(\Omega ))\quad \text {for}\ \frac{2}{p}+\frac{3}{q}-2=\frac{3}{5q}, \\{} & {} \begin{array}{c} \nabla g_1,\ \nabla g_2 \in L^\sigma (0,T;L^\tau (\Omega ))\quad \text {for}\ \dfrac{2}{\sigma } + \dfrac{3}{\tau }-2=\dfrac{3}{5\tau },\\ \text {with}\ \dfrac{9}{5}-\varepsilon _2(q) \leqq \tau < 3. \end{array} \end{aligned}$$ -
(ii-i)
We take \(\alpha \) and \(\beta \) so that
$$\begin{aligned} \frac{1}{q^*} \leqq \frac{1}{\alpha }<\frac{1}{2},\ \frac{1}{\tau ^*} \leqq \beta <\frac{1}{2} \text { and } \frac{1}{\tau '}=\frac{1}{\alpha }+\frac{1}{\beta }. \end{aligned}$$For such \(\alpha \) and \(\beta ,\) we take \(0 \leqq \theta ,\ \tilde{\theta }< 1\) so that
$$\begin{aligned} \frac{1}{\alpha }=\frac{\theta }{2}+\frac{1-\theta }{q^*},\ \frac{1}{\beta }=\frac{{{\tilde{\theta }}}}{2}+\frac{1-\tilde{\theta }}{\tau ^*}. \end{aligned}$$Then we have the estimate
$$\begin{aligned} \left| \int ^t_0 ((f \cdot \nabla )g_1,g_2) ds \right| \leqq C \Vert f\Vert ^{\theta }_{L^\infty L^2} \Vert \nabla f\Vert ^{1-\theta }_{L^p L^q} \Vert \nabla g_1\Vert _{L^\sigma L^\tau } \Vert g_2\Vert ^{{\tilde{\theta }}}_{L^\infty L^2} \Vert \nabla g_2\Vert ^{1 - \tilde{\theta }}_{L^\sigma L^\tau }, \end{aligned}$$where \(C = C(q,\tau ,\alpha ,\beta ).\)
-
(ii-ii)
We also take \(0 \leqq {\tilde{\theta }} < 1\) so that
$$\begin{aligned} \frac{1}{2q'} = \frac{{{\tilde{\theta }}}}{2} + \frac{1-\tilde{\theta }}{\tau ^*}. \end{aligned}$$Then we have the estimate
$$\begin{aligned} \left| \int ^t_0 ((g_1 \cdot \nabla ) f,g_2) ds \right| \leqq C \Vert g_1\Vert ^{{\tilde{\theta }}}_{L^\infty L^2} \Vert \nabla g_1\Vert ^{1 - {{\tilde{\theta }}}}_{L^\sigma L^\tau } \Vert \nabla f\Vert _{L^p L^q} \Vert g_2\Vert ^{{\tilde{\theta }}}_{L^\infty L^2} \Vert \nabla g_2\Vert ^{1 - \tilde{\theta }}_{L^\sigma L^\tau }, \end{aligned}$$where \(C = C(q,\tau ).\)
Proof of Lemma 3.1
In the case of (i-i), by the convex interpolation, we have
Now, take the exponents \(1 \leqq a_1,\ a_2,\ b_1,\ b_2 \leqq \infty \) as
Then we see that
Hence, it follows from the Hölder inequality in time that
In the case of (i-ii), by the convex interpolation, we have
Here, it should be noted that \(\tau ^* \leqq 2q' < 6\) holds by the assumptions \(3/2 < q\) and \(\tau \leqq 9/5+\varepsilon _1(q).\) Now, take the exponents \(1\leqq b_1,\) \(b_2 \leqq \infty \) as
Then we see that
Hence, the Hölder inequality gives
In the case of (ii-i), it follows from the convex interpolation that
Now, take the exponents \(1 \leqq a,\) \(b \leqq \infty \) as
Then we see that
Hence, the Hölder inequality gives
In the case of (ii-ii), it follows from the convex interpolation that
Here, it should be noted that \(2 < 2q' \leqq \tau ^*\) holds by the assumption \(9/5 - \varepsilon _2(q) \leqq \tau .\) Now, take the exponent \(1< b < \infty \) as
Then we see that
Hence, the Hölder inequality gives
This completes the proof of Lemma 3.1.
Now, we remark on the existence of the exponents \(\alpha \) and \(\beta \) in the proof. \(\square \)
Remark 2
1. Let \(3/2 < q \leqq 9/5\) and \(3/2 < \tau \leqq 9/5+\varepsilon _1(q).\) Let \(1/6 < 1/\beta \leqq { 1/\tau ^*}.\) Set \(\alpha \) as \(\tau ' \leqq \alpha \) and \(1/\tau '=1/\alpha +1/\beta .\) Then we have
where
Moreover, we can choose \(\beta \) satisfying \(1/6 < 1/\beta \leqq 1/\tau ^*\) and \(L_1 \leqq 1/\beta < L_2\) because we see that \((1/6, 1/\tau ^*] \cap [L_1, L_2) \ne \text{\O }\) by the assumption \(3/2 < \tau \leqq 9/5+\varepsilon _1(q).\)
2. Let \(9/5 \leqq q < 3\) and \(9/5- \varepsilon _2(q) \leqq \tau < 3.\) Let \(1/\tau ^* \leqq 1/\beta < 1/2.\) Set \(\alpha \) as \(\tau ' \leqq \alpha \) and \(1/\tau '=1/\alpha +1/\beta .\) Then we have
where
Moreover, we can choose \(\beta \) satisfying \(1/\tau ^* \leqq 1/\beta < 1/2\) and \(L_3 < 1/\beta \leqq L_1\) because we see that \([1/\tau ^*,1/2) \cap (L_3,L_1] \ne \text{\O }\) by the assumption \(9/5 - \varepsilon _2(q) \leqq \tau .\)
Lemma 3.2
Let \(3 \leqq q < \infty \) and \(f,\ g_1,\ g_2 \in L^\infty (0,T;L^2({\mathbb {R}}^3)) \cap L^2(0,T;H^1({\mathbb {R}}^3))\) with \(f \vert _{\partial \Omega } = g_1 \cdot \nu \vert _{\partial \Omega } = g_2 \cdot \nu \vert _{\partial \Omega } = 0.\) Let us define \(\Pi _{N,f,g_1,g_2}\) by
Suppose that f, \(g_1\) and \(g_2\) satisfy
(i) Case \(3 \leqq q < \infty \) and \(9/5-\varepsilon _2(q) < \tau \leqq 2 q'\): We take \(0< \rho < 1\) so that
For such \(\rho ,\) we take \(0< \alpha ,\ \beta < 1\) and \(0< {{\tilde{\rho }}} < 1\) so that
Then we have the estimate
where \(C = C(q,\tau ,\rho ).\)
(ii) Case \(3 \leqq q < \infty \) and \(2q'< \tau < \infty \): We take \(0< \rho < 1\) so that
For such \(\rho ,\) we take \(0 < \alpha ,\) \(\beta < 1\) and \(0< \tilde{\rho }< 1\) so that
Then we have the estimate
where \(C = C(q, \tau , \rho ).\)
We first prove a following proposition.
Proposition 3.3
Let f, \(g_1\) and \(g_2\) be as in Lemma 3.2. Let the exponents \(2 \leqq a,\ b,\ c < \infty \) and \(0< \alpha ,\ \beta ,\ \gamma < 1\) satisfy \(1=1/a+1/b+1/c\) and \(1=\alpha + \beta + \gamma ,\) respectively. Then we have
where C is a constant independent of \(N,\ f,\ g_1,\ g_2.\)
Proof of Proposition 3.3
Let us use the Bernstein inequality and imbedding \(B^0_{a,2} \hookrightarrow L^a\) to estimate
By the estimate (3.3), we have
Hence we obtain
Analogously, we have
\(\square \)
Proof of Lemma 3.2
In the case of (i), it follows from the Bernstein inequality that
Note the facts that
Hence, combining (3.4), (3.5) and Proposition 3.3, we have
where \(1<r<\infty \) is given by
By \(2r'{{\tilde{\rho }}}=\sigma ,\) the term \(\int ^t_0 \vert \Pi _{N,f,g_1,g_2} \vert ds\) is bounded by the RHS of the inequality (3.1). Similarly, we can also estimate the term \(\int ^t_0 \vert \Pi _{N,g_1,g_2,f} \vert ds.\) Hence, we obtain (3.1).
In the case of (ii), it follows from the Bernstein inequality that
We set \(0<{{\tilde{\theta }}} < 1\) as
Then, we have
Hence, combining (3.6), (3.7) and Proposition 3.3, we have
where the exponents r is given by
Here, note the fact
Then the Bernstein inequality gives
Combining (3.8) and (3.9), the term \(\int ^t_0 \vert \Pi _{N,f,g_1,g_2} \vert ds\) is bounded by the RHS of the inequality (3.2). Similarly, we can also estimate the term \(\int ^t_0 \vert \Pi _{N,g_1,g_2,f} \vert ds.\) Hence, we obtain (3.2). This completes the proof of Lemma 3.2. \(\square \)
4 Proof of main results
4.1 Proof of Theorem 2.1
Let \(X^{1,q}\) and \(Y^{1,\tau }\) be function spaces given by
In the same way as [1], let us take approximating functions \(\{u_m\}^\infty _{m=1}\) and \(\{B_m\}^\infty _{m=1}\) of u and B with \(\textrm{div}u_m = 0,\) \(u_m \vert _{\partial \Omega }=0,\) \(\textrm{div}B_m = 0\) and \(B_m \cdot \nu \vert _{\partial \Omega }=0\) in the norm of \(L^\infty (0,T;L^2(\Omega )) \cap L^2(0,T;H^1(\Omega )) \cap L^p(0,T;X^{1,q})\) and \(L^\infty (0,T; L^2(\Omega ))\cap L^2(0,T;H^1(\Omega )) \cap L^\sigma (0,T;Y^{1,\tau }),\) respectively. Let \(f^\varepsilon \) donate the mollifier of f in time, i.e.,
We take \((u_m)^\varepsilon \) and \((B_m)^\varepsilon \) as the test functions of the first and second equation in (1.1), respectively. Then adding the two identities, we obtain
where
For the proof of the energy equality (1.2), it suffices to show that
To show this, we first split \(h_1\) into following six terms:
Moreover, we can show that h(u, u, u) and h(u, B, B) are exactly zero since following facts holds by the same method as [1]:
In \(h_2,\) to control the error of the solution, we decompose \(h_2\) into the following forms:
Notice that all trilinear forms on the RHS of \(h_1\) and \(h_2\) in (4.1) and (4.2) are written in the following three types:
where
Here, by the virtue of (4.3) and (4.4), the following inequalities from the vector analysis hold:
Combining (4.5), (4.6) and Lemma 3.1, we see that the non-linear terms \(I_2\) and \(I_3\) are estimated in \(L^\infty (0,T;L^2(\Omega )) \cap L^2(0,T;H^1(\Omega )) \cap L^p(0,T;X^{1,q})\) and \(L^\infty (0,T;L^2(\Omega )) \cap L^2(0,T;H^1(\Omega )) \cap L^p(0,T;Y^{1,\tau }).\) Here, by \(\theta ,\ {{\tilde{\theta }}} \ne 1\) in the Lemma 3.1(ii), we can prove the convergence of the error term in \(L^2(0,T;H^1(\Omega )) \cap L^p(0,T;X^{1,q})\) and \(L^2(0,T;H^1(\Omega )) \cap L^p(0,T;Y^{1,\tau }).\) Notice that the terms \(\Vert f,f_i\Vert _{L^p L^q},\) \(i=1,2,3\) and \(\Vert g,g_i\Vert _{L^\sigma L^\tau },\) \(i=1,2\) appear because of RHS of (4.5) and (4.6). But it is easy to show that these terms are bounded by the norm of these functions in the Leray–Hopf class (see Remark 3 for details) . Moreover, it follows from the previous work [1] and (4.5) that the nonlinear term \(I_1\) is estimated in \(L^\infty (0,T;L^2(\Omega )) \cap L^2(0,T;H^1(\Omega )) \cap L^p(0,T;X^{1,q}).\) Hence, the solutions u and B satisfy the energy equality (1.2).
In the case \((0,T) \times {\mathbb {R}}^3,\) it should be noted that Theorem 2.1 is immediately established by the Lemma 3.1 because the terms \(\Vert f,f_i\Vert _{L^p L^q},\) \(i=1,2,3\) and \(\Vert g,g_i\Vert _{L^\sigma L^\tau },\) \(i=1,2\) in the RHS in (4.5) and (4.6) vanish. \(\Box \)
Remark 3
1. For \(2/r+3/s=3/2\) and \(0 < 1/r \leqq 1/2,\) the following imbedding holds
with the estimate
for some \(0 \leqq \gamma <1.\) Here, the constant C is independent of f. Note that \(0< 1/r \leqq 1/2\) is equivalent to \(1/6 \leqq 1/s <1/2.\)
2. For \(2/\sigma +3/\tau -2=2-3/\tau \) with \(1/2< 1/\tau < 2/3,\) it follows from (4.7) that
where \(2/\sigma +3/s = 3/2.\) Here, it should be noted that \(1/2< 1/\tau < 2/3\) is equivalent to \(0< 1/r = 1/\sigma <1/2.\)
3. For \(2/\sigma + 3/\tau -2 = 3/5\tau ,\) in the case \(5/12< 1/\tau < 2/3,\) it follows that
where \(2/\sigma +3/s = 3/2.\) Here, it should be noted that \(5/12< 1/\tau < 2/3\) is equivalent to \(1/5< 1/r = 1/\sigma <1/2.\) In the case \(1/3 < 1/\tau \leqq 5/12,\) it follows that
where \(2/r+3/\tau = 3/2.\) Here, it should be noted that \(1/6< 1/3< 1/\tau = 1/s< 5/12 < 1/2.\)
4. By \(1-\gamma \ne 0,\) we can prove the convergence of the error terms in \(L^2(0,T; H^1(\Omega )).\)
4.2 Proof of Theorem 2.2
In the same way as [4], we take \((u_{\leqq N})_{\leqq N}\) and \((B_{\leqq N})_{\leqq N}\) as the test functions of the first and second equality in (1.1), respectively. Then we obtain
By Lemma 3.2, we see that
Moreover, the estimate of the term \(\int ^t_0 \Pi _{N,u,u,u}ds\) was proved in [3]. Hence, u and B satisfy the energy equality (1.2). \(\square \)
Data availability statement
All data generated or analysed during this study are included in this published article.
References
Berselli, L.C., Chiodaroli, E.: On the energy equality for the 3D Navier–Stokes equations. Nonlinear Anal. 192, 111704 (2020). https://doi.org/10.1016/j.na.2019.111704
Caflisch, R.E., Klapper, I., Steele, G.: Remarks on singularities, dimension and energy dissipation for ideal hydrodynamics and MHD. Commun. Math. Phys. 184, 443–455 (1997). https://doi.org/10.1007/s002200050067
Cheskidov, A., Constantin, P., Friedlander, S., Shvydkoy, R.: Energy conservation and Onsager’s conjecture for the Euler equations. Nonlinearity 21, 1233–1252 (2008). https://doi.org/10.1088/0951-7715/21/6/005
Cheskidov, A., Luo, X.: Energy equality for the Navier–Stokes equations in weak-in-time Onsager spaces. Nonlinearity 33, 1388–1403 (2020). https://doi.org/10.1088/1361-6544/ab60d3
Constantin, P., E, W., Titi, E.S.: Onsager’s conjecture on the energy conservation for solutions of Euler equations. Commun. Math. Phys. 165, 207–209 (1994)
He, C., Xin, Z.: On the regularity of weak solutions to the magnetohydrodynamic equations. J. Differ. Equ. 213, 235–254 (2005). https://doi.org/10.1016/j.jde.2004.07.002
Hopf, E.: Über die Anfangswertaufgabe für die hydrodynamischen Grundgleichungen. Math. Nachr. 4, 213–231 (1951). https://doi.org/10.1002/mana.3210040121
Kang, E., Lee, J.: Remarks on the magnetic helicity and energy conservation for ideal magneto-hydrodynamics. Nonlinearity 20, 2681–2689 (2007). https://doi.org/10.1088/0951-7715/20/11/011
Leray, J.: Sur le mouvement d’un liquide visqueux emplissant l’espace. Acta Math. 63(1), 193–248 (1934). https://doi.org/10.1007/BF02547354
Serrin, J.: The initial value problem for the Navier–Stokes equations. In: Langer, R.E. (ed.) Nonlinear Problems, pp. 69–98. University of Wisconsin Press, Madison (1963)
Shinbrot, M.: The energy equation for the Navier–Stokes system. SIAM J. Math. Anal. 5, 948–954 (1974). https://doi.org/10.1137/0505092
Tan, W., Wu, F.: Energy conservation and regularity for the 3D magneto-hydrodynamics equations. Discret. Contin. Dyn. Syst. 42(11), 5487–5508 (2022). https://doi.org/10.3934/dcds.2022110
Taniuchi, Y.: On generalized energy equality of the Navier–Stokes equations. Manuscr. Math. 94, 365–384 (1997)
Wang, G., Zuo, B.: Energy equality for weak solutions to the 3D magnetohydrodynamic equations in a bounded domain. Discret. Contin. Dyn. Syst. Ser. B 27(2), 1001–1027 (2022). https://doi.org/10.3934/dcdsb.2021078
Wang, Y., Zuo, B.: Energy and cross-helicity conservation for the three-dimensional ideal MHD equations in a bounded domain. J. Differ. Equ. 268, 4079–4101 (2020). https://doi.org/10.1016/j.jde.2019.10.045
Yong, Y., Jiu, Q.: Energy equality and uniqueness of weak solutions to MHD equation in \(L^\infty (0, T;L^n(\Omega ))\). Acta. Math. Sin. Engl. Ser. 25, 803–814 (2009). https://doi.org/10.1007/s10114-009-7214-8
Zeng, Y.: Note on energy equality of MHD system. Z. Angew. Math. Phys. 73, 21 (2022). https://doi.org/10.1007/s00033-021-01653-0
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that there are no competing interests to declare.
Additional information
This article is part of the section “Theory of PDEs” edited by Eduardo Teixeira.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Eguchi, T. Energy equality of MHD system under a weaker condition on magnetic field. Partial Differ. Equ. Appl. 4, 39 (2023). https://doi.org/10.1007/s42985-023-00257-9
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s42985-023-00257-9
Keywords
- Magnetohydrodynamics equations
- Energy identity
- Energy conservation law
- Weak solutions
- Onsager’s conjecture