Discretely Self-Similar Solutions for 3D MHD Equations and Global Weak Solutions in Weighted \(L^2\) Spaces


This paper deals with the existence of global weak solutions for 3D MHD equations when the initial data belong to the weighted spaces \(L^2_{w_\gamma }\), with \(w_\gamma (x)=(1+\vert x\vert )^{-\gamma }\) and \(0 \le \gamma \le 2\). Moreover, we prove the existence of discretely self-similar solutions for 3D MHD equations for discretely self-similar initial data which are locally square integrable. Our methods are inspired of a recent work (Fernández-Dalgo et al. in Arch Rational Mech Anal 237:347–382, 2020) for the Navier–Stokes equations.

This is a preview of subscription content, access via your institution.


  1. 1.

    Alvarez-Samaniego, B., Alvarez-Samaniego, P., Fernández-Dalgo, P.: On the use of the Riesz transforms to determine the pressure term in the incompressible Navier–Stokes equations on the whole space. Preprint arXiv:2001.10436 (2020)

  2. 2.

    Basson, A.: Solutions spatialement homogènes adaptées des équations de Navier–Stokes. Université d’Évry, Thèse (2006)

  3. 3.

    Bradshaw, Z., Kukavica I., Tsai, T.P.: Existence of global weak solutions to the Navier–Stokes equations in weighted spaces. arXiv:1910.06929v1 (2019)

  4. 4.

    Bradshaw, Z., Tsai, T.P.: Global existence, regularity and uniqueness of infinite energy solutions to the Navier–Stokes equations. Commun. Partial Differ. Equ. 45, 1168–1201 (2020)

  5. 5.

    Bradshaw, Z., Tsai, T.P.: Discretely self-similar solutions to the Navier–Stokes equations with data in \(L^2_{{\rm loc}}\), to appear in Analysis and PDE (2019)

  6. 6.

    Chae, D., Wolf, J.: Existence of discretely self-similar solutions to the Navier–Stokes equations for initial value in \(L^2_{{\rm loc}}({\mathbb{R}}^3)\). Ann. Inst. Henri Poincaré Anal. Non Linéaire 35, 1019–1039 (2018)

    ADS  MathSciNet  Article  Google Scholar 

  7. 7.

    Fernández-Dalgo, P.G., Lemarié-Rieusset, P.G.: Weak solutions for Navier–Stokes equations with initial data in weighted \(L^2\) spaces. Arch. Rational Mech. Anal. 237, 347–382 (2020)

    ADS  MathSciNet  Article  Google Scholar 

  8. 8.

    Jia, H., Šverák, V.: Local-in-space estimates near initial time for weak solutions of the Navier–Stokes equations and forward self-similar solutions. Invent. Math. 196, 233–265 (2014)

    ADS  MathSciNet  Article  Google Scholar 

  9. 9.

    Kikuchi, N., Seregin, G.: Weak solutions to the Cauchy problem for the Navier–Stokes equations satisfying the local energy inequality. In: Birman, M.S., Uraltseva, N.N. (eds.) Nonlinear Equations and Spectral Theory. Amer. Math. Soc. Transl. Ser. 2, vol. 220, pp. 141–164. American Mathematical Soc., Providence (2007)

    Google Scholar 

  10. 10.

    Lemarié–Rieusset, P.G.: Solutions faibles d’énergie infinie pour les équations de Navier–Stokes dans \({\mathbb{R}}^{3}\). C. R. Acad. Sci. Paris, Serie I 328, 1133–1138 (1999)

    MathSciNet  Article  Google Scholar 

  11. 11.

    Lemarié-Rieusset, P.G.: Recent Developments in the Navier–Stokes Problem. CRC Press, Boca Raton (2002)

    Google Scholar 

  12. 12.

    Lemarié–Rieusset, P.G.: The Navier–Stokes Problem in the 21st Century. Chapman & Hall/CRC, Boca Raton (2016)

    Google Scholar 

  13. 13.

    Stein, E.M.: Harmonic Analysis, Real Variable Methods. Orthogonality and Oscillatory Integrals. Princeton University Press, New Jersey (1993)

    Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Pedro Gabriel Fernández-Dalgo.

Ethics declarations

Conflict of interest:

Pedro Gabriel Fernández-Dalgo and Oscar Jarrín declare that they have no conflict of interest.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Communicated by D. Chae

Approximated System

Approximated System

Let \(\theta \in {\mathcal {D}}({\mathbb {R}}^{3})\) be a non-negative, radial and radially decreasing function such that \(\int _{{\mathbb {R}}^{3}}\theta (x)dx=1\), and \(Supp (\theta ) \subset B(0,1)\). For \(\varepsilon >0\) we let \(\theta _\varepsilon (x) = \frac{1}{\varepsilon ^3} \theta (\frac{x}{\varepsilon })\). o

Proposition 1

Let \(\mathbf{u }_0 \in L^2({\mathbb {R}}^{3})\), \(\mathbf{b }_0 \in L^{2}({\mathbb {R}}^{3})\) be divergence free vector fields. Let \({\mathbb {F}} = (F_{i,j})_{1 \le i,j \le 2 } \) be a tensor such that \({\mathbb {F}} \in L^{2}((0, T), L^2)\), for all \(T< T_\infty \).

Then there exists a unique solution \((\mathbf{u }_\varepsilon , \mathbf{b }_\varepsilon , p_\varepsilon , q_\varepsilon )\) of the following approximated system

$$\begin{aligned} (MHD_\varepsilon ) \left\{ \begin{array}{ll} \partial _t \mathbf{u }= \Delta \mathbf{u }- [(\mathbf{u }*\theta _{\varepsilon }) \cdot \nabla ] \mathbf{u }+ [(\mathbf{b }*\theta _{\varepsilon }) \cdot \nabla ] \mathbf{b }- \nabla p + {{\varvec{\nabla }}}\cdot {\mathbb {F}}, \\ \partial _{t} \mathbf{b }= \Delta \mathbf{b }- [(\mathbf{u }_\varepsilon *\theta _{\varepsilon }) \cdot \nabla ] \mathbf{b }+ [(\mathbf{b }*\theta _\varepsilon ) \cdot \nabla ] \mathbf{u }- {{\varvec{\nabla }}}q ,\\ {{\varvec{\nabla }}}\cdot \mathbf{u }=0, \, {{\varvec{\nabla }}}\cdot \mathbf{b }=0, \\ \mathbf{u }(0,\cdot )=\mathbf{u }_0, \, \mathbf{b }(0,\cdot )=\mathbf{b }_0, \end{array} \right. \end{aligned}$$

on \([0,T_\infty )\) such that:

  1. (1)

    \(\mathbf{u }_\varepsilon , \mathbf{b }_\varepsilon \in L^{\infty } ([0, T),L^{2}({\mathbb {R}}^{3}))\cap L^{2}([0, T), {\dot{H}}^{1}({\mathbb {R}}^{3}))\), \( p_\varepsilon ,q_\varepsilon \in L^2((0,T), \dot{H} ^{-1}) +L^2((0,T), L^2) \), for all \(0<T < T_\infty \)

  2. (2)

    the pressure \(p_\varepsilon \) and the term \(q_\varepsilon \) are related to \(\mathbf{u }_\varepsilon , \mathbf{b }_\varepsilon \) and \({\mathbb {F}}\) by

    $$\begin{aligned} p_\varepsilon = \sum _{1 \le i,j\le 3} {\mathcal {R}}_i {\mathcal {R}}_j ((u_{\varepsilon , i}*\theta _\varepsilon ) u_{\varepsilon ,j} - (b_{\varepsilon ,i}*\theta _\varepsilon ) b _{\varepsilon ,j} - F_{i,j} ),\end{aligned}$$


    $$\begin{aligned} q_\varepsilon = \sum _{1 \le i,j\le 3} {\mathcal {R}}_i {\mathcal {R}}_j ( (u_{\varepsilon ,i}*\theta _\varepsilon ) b_{\varepsilon ,j} - (b_{\varepsilon ,j}*\theta _\varepsilon ) u _{\varepsilon ,i}), \end{aligned}$$

    where \({\mathcal {R}}_i =\frac{\partial _i}{\sqrt{-\Delta }}\) denote always the Riesz transforms. In particular, \( p_\varepsilon ,q_\varepsilon \in L^{4}((0,T),L^{6/5}) + L^{2}((0,T),L^{2})\).

  3. (3)

    The functions \((\mathbf{u }_\varepsilon , \mathbf{b }_\varepsilon , {\mathbb {F}}, {\mathbb {G}})\) verify the following global energy equality:

    $$\begin{aligned} \begin{aligned} \partial _t(\frac{\vert \mathbf{u }_\varepsilon \vert ^2 + |\mathbf{b }_\varepsilon |^2 }{2}) =&\, \Delta (\frac{\vert \mathbf{u }_\varepsilon \vert ^2 + |\mathbf{b }_\varepsilon |^2 }{2})-\vert {{\varvec{\nabla }}}\mathbf{u }_\varepsilon \vert ^2 - |{{\varvec{\nabla }}}\mathbf{b }_\varepsilon |^2 \\&- {{\varvec{\nabla }}}\cdot \left( (\frac{\vert \mathbf{u }_\varepsilon \vert ^2}{2} + \frac{\vert \mathbf{b }_\varepsilon \vert ^2}{2})(\mathbf{u }_\varepsilon * \theta _\varepsilon ) + p_\varepsilon \mathbf{u }_\varepsilon \right) \\&+ {{\varvec{\nabla }}}\cdot ((\mathbf{u }_\varepsilon \cdot \mathbf{b }_\varepsilon ) (\mathbf{b }_\varepsilon * \theta _\varepsilon ) + q_\varepsilon \mathbf{b }_\varepsilon ) \\&+ \mathbf{u }_\varepsilon \cdot ({{\varvec{\nabla }}}\cdot {\mathbb {F}}). \end{aligned} \end{aligned}$$


    $$\begin{aligned} \begin{aligned}&\Vert \mathbf{u }_\varepsilon (t) \Vert ^{2}_{L^2} + \Vert \mathbf{b }_\varepsilon (t) \Vert ^{2}_{L^2} + 2 \int _{a}^{t} (\Vert {{\varvec{\nabla }}}\mathbf{u }_\varepsilon (s) \Vert ^{2}_{L^2} + \Vert {{\varvec{\nabla }}}\mathbf{b }_\varepsilon (s) \Vert ^{2}_{L^2} ) ds \\&\quad = \Vert \mathbf{u }_\varepsilon (a) \Vert ^{2}_{L^2} + \Vert \mathbf{b }_\varepsilon (a) \Vert ^{2}_{L^2} - 2 \sum _{1 \le i,j \le 3} \int _{a}^{t} \int F_{i,j}\partial _iu_{_\varepsilon ,j}\ \, dx\, ds , \end{aligned} \end{aligned}$$

    which implies in particular

    $$\begin{aligned} \begin{aligned}&\Vert \mathbf{u }_\varepsilon (t) \Vert ^{2}_{L^2} + \Vert \mathbf{b }_\varepsilon (t) \Vert ^{2}_{L^2} + \int _{0}^{t} (\Vert {{\varvec{\nabla }}}\mathbf{u }_\varepsilon (s) \Vert ^{2}_{L^2} + \Vert {{\varvec{\nabla }}}\mathbf{b }_\varepsilon (s) \Vert ^{2}_{L^2} ) ds \\&\quad \le \Vert \mathbf{u }_0 \Vert ^{2}_{L^2} + \Vert \mathbf{b }_0 \Vert ^{2}_{L^2} + C \Vert {\mathbb {F}}\Vert ^{2}_{L^{2}_{t} L^{2}_{x}}. \end{aligned} \end{aligned}$$


We consider \(0< T< T_1 < T_\infty \) and the space \(E_T= {\mathcal {C}}([0,T], L^2({\mathbb {R}}^{3})) \cap L^2((0, T) {\dot{H}}^{1}({\mathbb {R}}^{3}))\) doted with the norm \(\Vert \cdot \Vert _{T}= \Vert \cdot \Vert _{L^{\infty }_{t}L^{2}_{x}}+ \Vert \cdot \Vert _{L^{2}_{t}{\dot{H}}^{1}_{x}}\). We will construct simultaneously \(\mathbf{u }_\varepsilon \) and \(\mathbf{b }_\varepsilon \). For this we will consider the space \(E_T \times E_T\) with the norm \(\Vert (\mathbf{u }_\varepsilon , \mathbf{b }_\varepsilon ) \Vert _{T}= \Vert \mathbf{u }_\varepsilon \Vert _{T}+ \Vert \mathbf{b }_\varepsilon \Vert _{T}\).

We use the Leray projection operator in order to express the problem \((MHD_\varepsilon )\) in terms of a fixed point problem. We let

$$\begin{aligned} a= e^{ t \Delta } (\mathbf{v }_0, \mathbf{c }_0)+ \int _{0}^{t} e^{(t-s)\Delta } {\mathbb {P}}( {{\varvec{\nabla }}}\cdot {\mathbb {F}}, (0,0,0) )(s,\cdot ) ds \end{aligned}$$


$$\begin{aligned} B((\mathbf{u },\mathbf{b }), (\mathbf{v },\mathbf{c })) = ( \, B_1((\mathbf{u },\mathbf{b })\, , \, (\mathbf{v },\mathbf{c })), B_2((\mathbf{u },\mathbf{b }), (\mathbf{v },\mathbf{c })) \, ) , \end{aligned}$$


$$\begin{aligned} B_1((\mathbf{u },\mathbf{b }), (\mathbf{v },\mathbf{c })) =&\int _{0}^{t} e^{(t-s)\Delta } {\mathbb {P}}([(\mathbf{u }*\theta _{\varepsilon }) \cdot \nabla ) \mathbf{v }- [(\mathbf{v }*\theta _{\varepsilon }] \cdot \nabla ] \mathbf{c })(s,\cdot ) ds ,\\ B_2((\mathbf{u },\mathbf{b }), (\mathbf{v },\mathbf{c })) =&\int _{0}^{t} e^{(t-s)\Delta } {\mathbb {P}}([(\mathbf{u }*\theta _{\varepsilon }) \cdot \nabla ] \mathbf{c }- [(\mathbf{b }*\theta _{\varepsilon }) \cdot \nabla ] \mathbf{v })(s,\cdot ) ds . \end{aligned}$$


$$\begin{aligned} (\mathbf{u }_\varepsilon , \mathbf{b }_\varepsilon , p_\varepsilon , q_\varepsilon ) \in E_T^2 \times \left( L^2((0,T), \dot{H} ^{-1}) +L^2((0,T), L^2) \right) ^2 \end{aligned}$$

is a solution of \((MHD_\varepsilon )\) if and only if \((\mathbf{u }_\varepsilon , \mathbf{b }_\varepsilon )\) is a fixed point for the application \((\mathbf{u },\mathbf{b }) \mapsto a + B((\mathbf{u },\mathbf{b }),(\mathbf{u },\mathbf{b }))\) and

$$\begin{aligned} p_\varepsilon = \sum _{1 \le i,j\le 3} {\mathcal {R}}_i {\mathcal {R}}_j ((u_{\varepsilon , i}*\theta _\varepsilon ) u_{\varepsilon ,j} - (b_{\varepsilon ,i}*\theta _\varepsilon ) b _{\varepsilon ,j} - F_{i,j} ) \end{aligned}$$


$$\begin{aligned} q_\varepsilon = \sum _{1 \le i,j\le 3} {\mathcal {R}}_i {\mathcal {R}}_j ( (u_{\varepsilon ,i}*\theta _\varepsilon ) b_{\varepsilon ,j} - (b_{\varepsilon ,j}*\theta _\varepsilon ) u _{\varepsilon ,i}) .\end{aligned}$$

We will use the Picard’s point fixed theorem. In order to study the linear terms, recall the following estimates, for a proof see [12], Theorem 12.2, page 352.

Lemma A.1

Let \(f \in L^{2}({\mathbb {R}}^{3})\) and \(g \in L^{2}_{t}{\dot{H}}^{-1}_{x}\). We have:

  1. (1)

    \(\Vert e^{ t \Delta } f \Vert _{T} \le c \Vert f \Vert _{L^2}\).

  2. (2)

    \(\left\| \int _{0}^{t} e^{(t-s)\Delta } g(s,\cdot )ds \right\| _{T} \le c(1+\sqrt{T}) \Vert g \Vert _{L^{2}_{t} {\dot{H}}^{-1}_{x}}\).

By this lemma we have

$$\begin{aligned} \Vert e^{ t \Delta } (\mathbf{u }_0, \mathbf{b }_0) \Vert _{T} \le c (\Vert \mathbf{u }_0 \Vert _{L^2} + \Vert \mathbf{b }_0 \Vert _{L^2}) \end{aligned}$$


$$\begin{aligned}&\left\| \int _{0}^{t} e^{(t-s)\Delta } {\mathbb {P}}( {{\varvec{\nabla }}}\cdot {\mathbb {F}}, (0,0,0))(s,\cdot ) ds \right\| _{T} \nonumber \\&\le c (1+\sqrt{T}) \Vert {\mathbb {P}}( {{\varvec{\nabla }}}\cdot {\mathbb {F}}) \Vert _{L^{2}_{t} {\dot{H}}^{-1}_{x}} \nonumber \\&\le c (1+\sqrt{T}) \Vert {\mathbb {F}}\Vert _{L^{2}_{t} L^{2}_{x}}. \end{aligned}$$

Now, to study the bilinear terms recall the following estimate given in [12] (Theorem 12.2, page 352):

Lemma A.2

Let \( \mathbf{u }, \mathbf{b }\in E_T\). We have

$$\begin{aligned} \begin{aligned} \displaystyle {\left\| \int _{0}^{t} e^{(t-s)\Delta } {\mathbb {P}}((( \mathbf{u }*\theta _{\varepsilon }) \cdot \nabla ) \mathbf{b })(s,\cdot )ds \right\| _{T} \le c \sqrt{T} \varepsilon ^{-3/2} \Vert \mathbf{u }\Vert _{T} \Vert \mathbf{b }\Vert _{T}}. \end{aligned} \end{aligned}$$

Applying this lemma to each bilinear term in the equation (22) we get

$$\begin{aligned} B((\mathbf{u },\mathbf{b }), (\mathbf{v },\mathbf{c })) \le c \sqrt{T}\varepsilon ^{-3/2} \Vert (\mathbf{u }, \mathbf{b }) \Vert _{T}\, \Vert (\mathbf{v }, \mathbf{c }) \Vert _{T}. \end{aligned}$$

Once we have inequalities (22), (23) and (24), for a time \(0< T_0 <T_1 \) such that

$$\begin{aligned} T_0 = \min \left( T_1, \frac{c \varepsilon ^3 }{(\Vert (\mathbf{u }_0, \mathbf{b }_0)\Vert _{L^2} + \Vert {\mathbb {F}}\Vert _{L^2((0,T_1), L^2)})^2} \right) , \end{aligned}$$

by the Picard’s contraction principle, we obtain \((\mathbf{u }_\varepsilon , \mathbf{b }_\varepsilon , p_\varepsilon , q_\varepsilon )\) a local solution of \((MHD_\varepsilon )\), where \( \mathbf{u }_\varepsilon , \mathbf{b }_\varepsilon \in E_T \) and \(p_\varepsilon , q_\varepsilon \in L^2((0,T), \dot{H} ^{-1}) +L^2((0,T), L^2)\). We can verify that this solution is unique.

Now, we prove that \( (\mathbf{u }_\varepsilon , \mathbf{b }_\varepsilon , p_\varepsilon , q_\varepsilon )\) is a global solution. We define the maximal existence time of the solution \(\mathbf{u }_\epsilon , \mathbf{b }_\epsilon \) by

$$\begin{aligned} T_{MAX} = \sup \{ 0< T \le T_\infty \, : \, \mathbf{u }_\epsilon , \mathbf{b }_\epsilon \in E_T \} \end{aligned}$$

If \( T_{MAX} < T_{\infty }\) we take \( 0< T< T_{MAX}< T_1 < T_{\infty }\), then \((\mathbf{u }_\epsilon , \mathbf{b }_\epsilon )\) is a solution of (\(MHD_\varepsilon \)) on [0, T] and \((\mathbf{u }_\epsilon , \mathbf{b }_\epsilon )\) is a solution on \([T, T+ \delta ]\), where

$$\begin{aligned} \delta = \min \left( T_1 - T , \frac{c \varepsilon ^3 }{(\Vert (\mathbf{u }_\epsilon (T), \mathbf{b }_\epsilon (T))\Vert _{L^2} + \Vert {\mathbb {F}}\Vert _{L^2((T,T_1), L^2)} )^2} \right) , \end{aligned}$$

which implies that \(\lim _{T \rightarrow T_{MAX}^{-} } \Vert (\mathbf{u }_\varepsilon (T), \mathbf{b }_\varepsilon (T) ) \Vert _{L^2} = + \infty \), however, we will see that it is not possible.

As \( ((\mathbf{b }_{\varepsilon } * \theta _\varepsilon ) \cdot {{\varvec{\nabla }}}) \mathbf{b }_{\varepsilon } ) \mathbf{u }_{_\varepsilon } = {{\varvec{\nabla }}}\cdot ( \mathbf{b }_{\varepsilon } \otimes ( \mathbf{b }_{\varepsilon } * \theta _\varepsilon )) \mathbf{u }_{_\varepsilon }\) belongs to \(L^2((0,T), \dot{H} ^{-1})\), and the same for the other non linear terms, we can write

$$\begin{aligned} \frac{d}{dt} \Vert \mathbf{u }_\varepsilon (t) \Vert ^{2}_{L^2}= & {} 2 \langle \partial _t \mathbf{u }_\varepsilon (t), \mathbf{u }_\varepsilon (t) \rangle _{{\dot{H}}^{-1}\times {\dot{H}}^{1}} \\= & {} -2 \Vert {{\varvec{\nabla }}}\mathbf{u }_\varepsilon (t) \Vert ^{2}_{L^2} + 2 \sum _{1\le i,j \le 3} \int b_{\varepsilon ,i} ( b_{\varepsilon ,j} * \theta _\varepsilon ) \partial _i u_{_\varepsilon ,j} dx \\&- 2 \sum _{1\le i,j \le 3} \int F_{i,j}\partial _iu_{_\varepsilon ,j}\ \, dx , \end{aligned}$$


$$\begin{aligned} \frac{d}{dt} \Vert \mathbf{b }_\varepsilon (t) \Vert ^{2}_{L^2}= & {} 2 \langle \partial _t \mathbf{b }_\varepsilon (t), \mathbf{b }_\varepsilon (t) \rangle _{{\dot{H}}^{-1}\times {\dot{H}}^{1}} \\= & {} -2 \Vert {{\varvec{\nabla }}}\mathbf{b }_\varepsilon (t) \Vert ^{2}_{L^2} + 2 \sum _{1\le i,j \le 3} \int u_{\varepsilon ,i} ( b_{\varepsilon ,j} * \theta _\varepsilon ) \partial _i b_{_\varepsilon ,j} dx . \end{aligned}$$

where we have used the fact that

$$\begin{aligned} \int ( (\mathbf{u }_\varepsilon * \theta ) \cdot \nabla ) \mathbf{b }_\varepsilon \cdot \mathbf{b }_\varepsilon \, dx&= \int \sum _{1\le i,j \le 3} ( (u_{j, _\varepsilon } *\theta ) \partial _j b_{i,_\varepsilon }) b_{ i, _\varepsilon }\, dx \\&= -\frac{1}{2} \int (\mathbf{u }_{ \varepsilon } *\theta ) \cdot {{\varvec{\nabla }}}( |\mathbf{b }_\varepsilon |^2 ) \, dx \\&=-\frac{1}{2} \int \nabla \cdot (\mathbf{u }_\varepsilon *\theta _\varepsilon ) |\mathbf{b }_\varepsilon |^2 dx = 0. \end{aligned}$$

Then, an integration by parts gives

$$\begin{aligned} \sum _{1\le i,j \le 3} \int u_{\varepsilon ,i} ( b_{\varepsilon ,j} * \theta _\varepsilon ) \partial _i b_{\varepsilon ,j} dx = - \sum _{1\le i,j \le 3} \int b_{\varepsilon ,i} ( b_{\varepsilon ,j} * \theta _\varepsilon ) \partial _i u_{\varepsilon ,j} dx, \end{aligned}$$

so we have

$$\begin{aligned} \frac{d}{dt} ( \Vert \mathbf{u }_\varepsilon (t) \Vert ^{2}_{L^2}+ \Vert \mathbf{b }_\varepsilon (t) \Vert ^{2}_{L^2})= & {} -2 (\Vert {{\varvec{\nabla }}}\mathbf{u }_\varepsilon (t) \Vert ^{2}_{L^2} + \Vert {{\varvec{\nabla }}}\mathbf{b }_\varepsilon (t) \Vert ^{2}_{L^2} ) \\&- 2 \sum _{1 \le i,j \le 3} \int F_{i,j}\partial _iu_{j}\ \, dx\, ds . \end{aligned}$$

By integrating on the time interval [0, T] we obtain the control (20) which implies by Grönwall inequality that \(\Vert (\mathbf{u }_\varepsilon , \mathbf{b }_\varepsilon )(T) ) \Vert _{L^2}\) does not converges to \(+\infty \) when T goes to \(T_{MAX}\) if \(T_{MAX}< T_\infty \), hence the solution is defined on \([0, T_\infty )\). Finally, remark that we can write

$$\begin{aligned} {{\varvec{\nabla }}}\cdot ((\mathbf{b }_\varepsilon \cdot \mathbf{u }_\varepsilon ) (\mathbf{b }_\varepsilon *\theta _\varepsilon ))&= {{\varvec{\nabla }}}(\mathbf{b }_\varepsilon \cdot \mathbf{u }_\varepsilon ) \cdot (\mathbf{b }_\varepsilon *\theta _\varepsilon ) \\&=((\mathbf{b }_\varepsilon *\theta _\varepsilon ) \cdot {{\varvec{\nabla }}}) \mathbf{b }_\varepsilon \cdot \mathbf{u }_\varepsilon + ((\mathbf{b }_\varepsilon *\theta _\varepsilon ) \cdot {{\varvec{\nabla }}}) \mathbf{u }_\varepsilon ) \cdot \mathbf{b }_\varepsilon \end{aligned}$$

so that

$$\begin{aligned} \partial _t(\frac{\vert \mathbf{u }_\varepsilon \vert ^2}{2})=&\, \Delta (\frac{\vert \mathbf{u }_\varepsilon \vert ^2}{2})-\vert {{\varvec{\nabla }}}\mathbf{u }_\varepsilon \vert ^2- {{\varvec{\nabla }}}\cdot \left( \frac{\vert \mathbf{u }_\varepsilon \vert ^2}{2} (\mathbf{u }_\varepsilon * \theta _\varepsilon ) + p_\varepsilon \mathbf{u }_\varepsilon \right) \\&+ {{\varvec{\nabla }}}\cdot ((\mathbf{u }_\varepsilon \cdot \mathbf{b }_\varepsilon ) (\mathbf{b }_\varepsilon * \theta _\varepsilon )) - ((\mathbf{b }_\varepsilon *\theta _\varepsilon ) \cdot {{\varvec{\nabla }}}) \cdot \mathbf{u }_\varepsilon ) \mathbf{b }_\varepsilon + \mathbf{u }_\varepsilon \cdot ({{\varvec{\nabla }}}\cdot {\mathbb {F}}), \end{aligned}$$

similarly we find

$$\begin{aligned} \partial _t(\frac{\vert \mathbf{b }_\varepsilon \vert ^2}{2})=&\, \Delta (\frac{\vert \mathbf{b }_\varepsilon \vert ^2 }{2})-\vert {{\varvec{\nabla }}}\mathbf{b }_\varepsilon \vert ^2 -{{\varvec{\nabla }}}\cdot \left( \frac{\vert \mathbf{b }_\varepsilon \vert ^2}{2}(\mathbf{u }_\varepsilon * \theta _\varepsilon ) + q_\varepsilon \mathbf{b }_\varepsilon \right) \\&+ ((\mathbf{b }_\varepsilon *\theta _\varepsilon ) \cdot {{\varvec{\nabla }}}) \cdot \mathbf{u }_\varepsilon ) \mathbf{b }_\varepsilon . \end{aligned}$$

By adding these equations we obtain the energy equality (19). \(\square \)

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Fernández-Dalgo, P.G., Jarrín, O. Discretely Self-Similar Solutions for 3D MHD Equations and Global Weak Solutions in Weighted \(L^2\) Spaces. J. Math. Fluid Mech. 23, 22 (2021). https://doi.org/10.1007/s00021-020-00551-1

Download citation


  • MHD equations
  • Weighted \(L^2\) spaces
  • Discretely self-similar solutions
  • Energy controls

Mathematics Subject Classification

  • 35Q30
  • 76D05