Skip to main content

On the regularity criterion for the 3D generalized MHD equations in Besov spaces

Abstract

In this paper, we consider the three-dimensional generalized MHD equations, a system of equations resulting from replacing the Laplacian Δ in the usual MHD equations by a fractional Laplacian ( Δ ) α . We obtain a regularity criterion of the solution for the generalized MHD equations in terms of the summation of the velocity field u and the magnetic field b by means of the Littlewood-Paley theory and the Bony paradifferential calculus, which extends the previous result.

MSC: 76B03, 76D03.

1 Introduction

In this paper, we are concerned with the following three-dimensional generalized MHD equations:

{ u t + u u b b + ( Δ ) α u + P = 0 , b t + u b b u + ( Δ ) β b = 0 , u = b = 0 , u ( x , 0 ) = u 0 ( x ) , b ( x , 0 ) = b 0 ( x ) , x R 3 ,t>0,
(1)

where u(x,t) denotes the fluid velocity vector field, b is the magnetic field, P=P(x,t) is the scalar pressure; while u 0 (x) and b 0 (x) are the given initial velocity and initial magnetic fields, respectively, in the sense of distributions, with u 0 = b 0 =0; α,β>0 are the parameters. The generalized MHD equations generalize the usual MHD equations by replacing the Laplacian Δ by a general fractional Laplacian ( Δ ) α , ( Δ ) β . As α=β=1, the generalized MHD equations reduce to the usual MHD equations; when α=β=1 and b=0, the generalized MHD equations reduce to the Navier-Stokes equations. Moreover, it has similar scaling properties and energy estimate to the Navier-Stokes equations and the MHD equations. The study of system (1) will improve our understanding of the Navier-Stokes equations and the MHD equations.

For the 3D generalized MHD equations (1), Wu [1] showed that the system (1) possesses a global weak solutions corresponding to any L 2 initial data. Yet, just like the 3D Navier-Stokes equations and the 3D MHD equations, whether there exists a global smooth solution for the 3D generalized MHD equations (1) or not is an open problem. Recently, many authors studied the regularity problem for the 3D generalized MHD equations (1) intensively. Wu [1], [2] obtained some regularity criteria only relying on the velocity u. Zhou [3] considered the following two cases: 1α=β 3 2 and 1β 5 4 α< 5 2 and established the Serrin-type criteria involving velocity u. Wu [4] obtained the classical Beal-Kato-Majda criterion for the system (1). By means of the Fourier localization technique and the Bony paraproduct decomposition, Yuan [5] extended the Serrin-type criterion to

u L q ( 0 , T ; B p , s ( R 3 ) ) ,
(2)

with 2 α q + 3 p 2α1+s, 3 2 α 1 + s <p, 1<s1, (p,s)(,1) provided that 1α=β 5 4 .

On the other hand, suggested by the results in [6], [7], one may presume upon that there should have some cancelation properties between the velocity field u and the magnetic field b. When α=β, plus and minus the first equation of (1) and the second one, respectively, the system (1) can be rewritten as

{ W + t + W W + + ( Δ ) α W + + P = 0 , W t + W + W + ( Δ ) α W + P = 0 , W + = W = 0 , W + ( x , 0 ) = W 0 + ( x ) , W ( x , 0 ) = W 0 ( x ) ,
(3)

where

W ± =u±b, W 0 ± (x)= u 0 (x)± b 0 (x).

In this paper, we are interested in what kind of confluence the integrability of W + or W brings to the weak solution (u,b) of the system (1). Furthermore, we shall make efforts to establish some new regularity criterion of weak solutions of the system (3) in terms of W + or W . When α=β=1, He and Wang [8], Gala [9], Dong et al.[10] establish some regularity criteria in Lorentz spaces, the multiplier space, and the nonhomogeneous Besov space, respectively.

The purpose of this paper is to deal with the case α=β of the system (3), to establish certain kind of regularity criteria. The tools we use here are the Littlewood-Paley theory and the Bony paraproduct decomposition. Before stating our main result, we firstly recall the definition of weak solutions to the 3D generalized MHD equations (1) as follows.

Definition 1.1

Suppose that u 0 , b 0 L 2 ( R 3 ). The vector-valued function (u,b) is called a weak solution to the system (1) on R 3 ×(0,T), if it satisfies the following properties:

  1. (1)

    u L (0,T; L 2 ( R 3 )) L 2 (0,T; H α ( R 3 )), b L (0,T; L 2 ( R 3 )) L 2 (0,T; H β ( R 3 ));

  2. (2)

    u=0 and b=0 in the sense of a distribution;

  3. (3)

    for any ϕ,φ C 0 ( R 3 ×[0,T)) with ϕ=0 and φ=0, one has

    0 T R 3 ( ϕ t + u ϕ ) udxdt+ R 3 u 0 ϕ(x,0)dx= 0 T R 3 ( u Λ 2 α ϕ + b ϕ b ) dxdt,

and

0 T R 3 ( φ t + u φ ) bdxdt+ R 3 b 0 φ(x,0)dx= 0 T R 3 ( b Λ 2 β φ + b φ u ) dxdt,

where Λ= ( Δ ) 1 2 .

The weak solution ( W + , W ) to the system (3) can be defined in a similar way as follows.

Definition 1.2

Suppose that W 0 + , W 0 L 2 ( R 3 ). The vector-valued function ( W + , W ) is called a weak solution to the system (3) on R 3 ×(0,T), if it satisfies the following properties:

  1. (1)

    W + , W L (0,T; L 2 ( R 3 )) L 2 (0,T; H α ( R 3 ));

  2. (2)

    W + =0 and W =0 in the sense of a distribution;

  3. (3)

    for any ϕ,φ C 0 ( R 3 ×[0,T)) with ϕ=0 and φ=0, one has

    0 T R 3 ( W + ϕ t + ϕ : ( W W + ) + W + Λ 2 α ϕ ) d x d t + R 3 W 0 + ϕ ( x , 0 ) d x = 0

and

0 T R 3 ( W φ t + φ : ( W + W ) + W Λ 2 α φ ) d x d t + R 3 W 0 φ ( x , 0 ) d x = 0 ,

where Λ= ( Δ ) 1 2 .

From the above two definitions as regards weak solutions, the systems (1) and (3) are equivalent. It is easy to see that ( W + , W ) also verifies the system (3) in the sense of distribution, provided that (u,b) is a weak solution to the system (1) as α=β.

Now our main result can be stated.

Theorem 1.3

Suppose that1α=β 5 4 , the initial velocity and magnetic field( u 0 , b 0 ) H 1 ( R 3 )and u 0 = b 0 =0in the sense of distribution. Assume that(u,b)is a weak solution to the system (1) on some interval[0,T]with0<T. If W + satisfies the following condition:

W + L q ( 0 , T ; B p , s ( R 3 ) ) ,
(4)

with 2 α q + 3 p 2α1+s, 3 2 α 1 + s <p, 1<s1, (p,s)(,1). Then the weak solution(u,b)remains smooth on R 3 ×(0,T].

Remark 1.4

It should be mentioned that our condition (4) on W + does not seem comparable with (2) on u at least there is no inclusion between them.

Remark 1.5

Our result here improves the recent result obtained by Dong et al.[10] as α=β=1 except that s=1. However, the method of [10] cannot also be applied to the case s=1 in this paper. We shall consider this problem in the future.

Notation

Throughout the paper, C stands for generic constant. We use the notation AB to denote the relation ACB, and the notation AB to denote the relations AB and BA. For convenience, given a Banach space X, we denote its norm by X . If there is no ambiguity, we omit the domain of function spaces.

This paper is structured as follows. In Section 2, we introduce the Littlewood-Paley decomposition and the Bony paradifferential calculus. In Section 3, we give the proof of Theorem 1.3 by means of the Littlewood-Paley theory and the Bony paradifferential calculus.

2 The Littlewood-Paley theory

In this section, we will provides the Littlewood-Paley theory and the related facts.

Let S( R 3 ) be the Schwarz functions of rapidly decreasing functions. Given fS( R 3 ), its Fourier transformation Ff= f ˆ is defined by

f ˆ (ξ)= ( 2 π ) 3 2 R 3 e i x ξ f(x)dx.

Take two nonnegative radial functions X,ψS( R 3 ) supported respectively in B={ξ R 3 ,|ξ| 4 3 } and C={ξ R 3 , 3 4 |ξ| 8 3 } such that

X(ξ)+ j 0 ψ ( 2 j ξ ) =1,ξ R 3 .
(5)

Let h= F 1 ψ and h ˜ = F 1 X. The frequency localization operator is defined by

Δ j f = ψ ( 2 j D ) f = 2 3 j R 3 h ( 2 j y ) f ( x y ) d y , j 0 , S j f = X ( 2 j D ) f = 1 k j 1 Δ k f = 2 3 j R 3 h ˜ ( 2 j y ) f ( x y ) d y , Δ 1 f = S 0 f , Δ j f = 0 for  j 2 .

Formally, Δ j is a frequency projection to the annulus {|ξ| 2 j }, and S j is a frequency projection to the ball {|ξ| 2 j }. The above dyadic decomposition has a nice quasi-orthogonality, with the choice of X and ψ; namely, for any f,gS( R 3 ), we have the following properties:

Δ i Δ j f 0 , | i j | 2 , Δ i ( S j 1 f Δ j g ) 0 , | i j | 5 .
(6)

Details of the Littlewood-Paley decomposition theory can be found in [11], [12].

Let sR, the homogeneous Sobolev space is defined by

H s ( R 3 ) = { f S ( R 3 ) ; f H s < + } ,

where

f H s 2 = j = 1 2 2 j s Δ j f L 2 2

and the set S ( R 2 ) of temperate distributions is the dual set of S for the usual pairing.

When dealing with our problems, we will use some paradifferential calculus [12], [13]. It is a nice way to define a generalized product between temperate distributions, which is continuous in fractional Sobolev spaces, and which yet does not make any sense for the usual product. Let f, g be two temperate distributions. We denote

T f g i j Δ i f Δ j g= j S j 1 f Δ j g,R(f,g) | i j | 2 Δ i f Δ j g.
(7)

At least, we have the following the Bony decomposition:

fg= T f g+ T g f+R(f,g),
(8)

where the paraproduct T is a bilinear continuous operator. For simplicity, we denote

T f g= T f g+R(f,g).

We now introduce the inhomogeneous Besov spaces.

Definition 2.1

Let sR, 1p,q. The inhomogeneous Besov space B p , q s ( R 3 ) is defined by

B p , q s ( R 3 ) = { f S ( R 3 ) ; f B p , q s < } ,

where

f B p , q s ={ ( j = 1 2 j s q Δ j f L p q ) 1 q , for  q < , sup j 1 2 j s Δ j f L p , for  q = .

The following lemmas will be useful in the proof of our main result.

Lemma 2.2

(Bernstein Inequality, [13])

Let1pq. Assume thatf L p ( R 3 ), then there exists a constantCindependent off, jsuch that

supp f ˆ { ξ : | ξ | 2 j } f L p C 2 j | α | sup | β | = | α | β f L p , supp f ˆ { ξ : | ξ | 2 j } α f L q C 2 j | α | + 3 j ( 1 p 1 q ) f L p .

Lemma 2.3

(Embedding Results, [14])

  1. (1)

    Let 1p, 1q, q 1 , and s s 1 >0. Assume that either s> s 1 or s= s 1 and q q 1 . Then B p , q s ( R 3 ) B q , q 1 s 1 ( R 3 ).

  2. (2)

    If 1p p 1 and s= s 1 +3( 1 p 1 p 1 ), then B p , q s ( R 3 ) B p 1 , q s 1 ( R 3 ).

3 Proof of Theorem 1.3

This section is devoted to the proof of Theorem 1.3. Since B p , 2 α q + 3 p ( 2 α 1 ) ( R 3 ) B , 2 α q ( 2 α 1 ) ( R 3 ) by Lemma 2.3, we only need to prove that Theorem 1.3 holds in the case p=, that is, to prove that if W + B , s ( R 3 ) satisfies 2 α q 2α1+s, then the weak solution (u,b) to the system (1) is regular on R 3 ×(0,T].

Now, we denote u k = Δ k u, θ k = Δ k θ, P k = Δ k P, and take γ= 2 α 2 α 1 + s . Then we have s= 2 α γ (2α1). Applying the operator Δ k to both sides of (3), we get

W k + t + Δ k ( W W + ) + ( Δ ) α W k + + P k = 0 , W k t + Δ k ( W + W ) + ( Δ ) α W k + P k = 0 .
(9)

Then multiplying the first and the second equation of (9) by W k + , W k , respectively, and by Lemma 2.2, we obtain for k0

1 2 d d t W k + L 2 2 + c 2 2 k α W k + L 2 2 = Δ k ( W W + ) , W k + , 1 2 d d t W k L 2 2 + c 2 2 k α W k L 2 2 = Δ k ( W + W ) , W k .
(10)

Denote Θ k (t) ( W k + ( t ) L 2 2 + W k ( t ) L 2 2 ) 1 2 . Adding the two equations in (10), we have

1 2 d d t Θ k 2 (t)+c 2 2 k α Θ k 2 (t)= Δ k ( W W + ) , W k + Δ k ( W + W ) , W k .
(11)

Note that

W W + , W + =0, W + W , W =0.

Then (11) can be rewritten as

1 2 d d t Θ k 2 ( t ) + c 2 2 k α Θ k 2 ( t ) = [ W , Δ k ] W + , W k + + [ W + , Δ k ] W , W k I + II ,
(12)

where the commutator operator [A,B]=ABBA. By the Bony decomposition (8), the term I can be rewritten as

I = [ T ( W ) i , Δ k ] i W + , W k + + T Δ k i W + ( W ) i , W k + Δ k ( T i W + ( W ) i ) , W k + Δ k ( R ( ( W ) i , i W + ) ) , W k + I 1 + I 2 + I 3 + I 4 .

From the definition of Δ k , we have

[ T ( W ) i , Δ k ] i W + = | k k | 4 [ S k 1 ( W ) i , Δ k ] i W k + = | k k | 4 [ S k 1 ( ( W ) i ) Δ k i W k + Δ k ( S k 1 ( ( W ) i ) i W k + ) ] = | k k | 4 2 3 k R 3 h ( 2 k ( x y ) ) [ S k 1 ( W ) i ( x ) S k 1 ( W ) i ( y ) ] i W k + ( y ) d y = | k k | 4 2 4 k R 3 0 1 y S k 1 ( W ) i ( x τ y ) d τ i h ( 2 k y ) W k + ( x y ) d y .

Then by the Minkowski inequality, we get

| I 1 | W k + L 2 | k k | 4 S k 1 W L 2 W k + L k k 2 | k k | 4 2 k W k L 2 W k + L 2 W k + L k k 2 | k k | 4 2 k 2 k ( 2 α 1 2 α γ ) Θ k Θ k W + B , s k k 2 2 k 2 k ( 2 α 1 2 α γ ) Θ k Θ k W + B , s .

For the term I 2 ,

T Δ k i W + ( W ) i = j S j 1 ( Δ k i W + ) Δ j ( W ) i + | j j | 1 Δ j ( Δ k i W + ) Δ j ( W ) i = j ( 1 j j 2 Δ j ( Δ k i W + ) + ( Δ j 1 ( Δ k i W + ) + Δ j ( Δ k i W + ) + Δ j + 1 ( Δ k i W + ) ) ) Δ j ( W ) i = j ( 1 j j + 1 Δ j ( Δ k i W + ) ) Δ j ( W ) i = k k 2 S k + 2 ( Δ k i W + ) Δ k ( W ) i ,

where j is replaced by k in the last equality. Noticing that S k + 2 Δ k W + = Δ k W + for k >k, then by Lemma 2.2 and the Minkowski inequality, we have

| I 2 | = | k k 2 S k + 2 Δ k i W + Δ k ( W ) i , W k + | | k k | 2 | S k + 2 Δ k i W + Δ k ( W ) i , W k + | + k k 2 | S k + 2 Δ k i W + Δ k ( W ) i , W k + | W k + L 2 | k k | 2 S k 1 W + L W k L 2 + 2 k W k + L k k 2 W k + L 2 W k L 2 k k 2 | k k | 2 2 2 k α ( 1 1 γ ) Θ k Θ k W + B , s + k k 2 2 2 k α ( 1 1 γ ) Θ k 2 W + B , s k k 2 2 2 k α ( 1 1 γ ) Θ k 2 W + B , s + k k 2 2 2 k α ( 1 1 γ ) Θ k 2 W + B , s .

Using the support of the Fourier transformation of the term T i W + ( W ) i , we get

Δ k ( T i W + ( W ) i ) = Δ k ( k S k 1 ( i W + ) Δ k ( W ) i ) = | k k | 4 Δ k ( S k 1 ( i W + ) ( W k ) i ) .

By the Minkowski inequality and Lemma 2.2, we have

| I 3 | | k k | 4 Δ k ( S k 1 ( i W + ) ( W k ) i ) L 2 W k + L 2 W k + L 2 | k k | 4 S k 1 W + L W k L 2 k k 2 | k k | 4 2 2 k α ( 1 1 γ ) Θ k Θ k W + B , s k k 2 2 2 k α ( 1 1 γ ) Θ k 2 W + B , s .

With the incompressibility condition u=0, we have

Δ k R ( ( W ) i , i W + ) = Δ k ( | k k | 1 Δ k ( W ) i Δ k ( i W + ) ) = k , k k 2 ; | k k | 1 Δ k ( Δ k ( W ) i Δ k ( i W + ) ) = k , k k 2 ; | k k | 1 i Δ k ( Δ k ( W ) i Δ k W + ) ,

then by Lemma 2.2, we get

| I 4 | 2 k W k + L k k 2 W k + L 2 W k L 2 k k 2 2 2 k α ( 1 1 γ ) Θ k 2 W + B , s .

Combining the above estimates for I 1 - I 4 , we have

| I | k k 2 2 k 2 k ( 2 α 1 2 α γ ) Θ k Θ k W + B , s + k k 2 2 2 k α ( 1 1 γ ) Θ k 2 W + B , s + k k 2 2 2 k α ( 1 1 γ ) Θ k 2 W + B , s .
(13)

For the second term II of (12), by similar arguments to the ones used to derive (13) one can get

| II | k k 2 2 2 k α ( 1 1 γ ) Θ k 2 W + B , s + k k 2 2 k 2 k ( 2 α 1 2 α γ ) Θ k Θ k W + B , s + k k 2 2 k 2 k ( 2 α 1 2 α γ ) Θ k Θ k W + B , s .
(14)

Inserting (13) and (14) into (12) one infers that

1 2 d d t Θ k 2 ( t ) + 2 2 k α Θ k 2 ( t ) k k 2 2 2 k α ( 1 1 γ ) Θ k 2 W + B , s + k k 2 2 2 k α ( 1 1 γ ) Θ k 2 W + B , s + k k 2 2 k 2 k ( 2 α 1 2 α γ ) Θ k Θ k W + B , s + k k 2 2 k 2 k ( 2 α 1 2 α γ ) Θ k Θ k W + B , s .
(15)

The proof of Theorem 1.3 in the rest of this section is divided into two cases.

Case I: s[0,1). Take σ(0, 1 α ), then 1σα>0. Let

φ(t) sup k 1 2 k σ α Θ k (t),ω(t) sup k 1 2 2 k α ( σ + 1 ) 0 t Θ k 2 (τ)dτ.

Multiplying (15) with 2 2 k σ α and integrating with respect to t, we have

2 2 k σ α Θ k 2 ( t ) 2 2 k σ α Θ k 2 ( 0 ) + 2 2 k α ( σ + 1 ) 0 t Θ k 2 ( τ ) d τ 0 t k k 2 2 2 k σ α 2 2 k α ( 1 1 γ ) Θ k 2 W + B , s d τ + 0 t k k 2 2 2 k σ α 2 2 k α ( 1 1 γ ) Θ k 2 W + B , s d τ + 0 t k k 2 2 2 k σ α 2 k 2 k ( 2 α 1 2 α γ ) Θ k Θ k W + B , s d τ + 0 t k k 2 2 2 k σ α 2 k 2 k ( 2 α 1 2 α γ ) Θ k Θ k W + B , s d τ A 1 + A 2 + A 3 + A 4 .
(16)

By Lemma 2.2 and the Hölder inequality, we have

A 1 = 0 t k k 2 ( 2 k σ α Θ k ) 2 γ ( 2 k ( σ + 1 ) α Θ k ) 2 2 γ 2 ( k k ) ( 2 2 γ ) W + B , s d τ 0 t ( sup k 1 2 k σ α Θ k ) 2 γ ( sup k 1 2 k α ( σ + 1 ) Θ k ) 2 2 γ W + B , s d τ ( 0 t W + ( τ ) B , s γ φ 2 ( τ ) d τ ) 1 γ ω 1 1 γ ( t ) , A 2 = 0 t k k 2 ( 2 k σ α Θ k ) 2 γ ( 2 k ( σ + 1 ) α Θ k ) 2 2 γ 2 2 α ( k k ) ( σ + 1 2 γ ) W + B , s d τ 0 t ( sup k 1 2 k σ α Θ k ) 2 γ ( sup k 1 2 k α ( σ + 1 ) Θ k ) 2 2 γ W + B , s d τ ( 0 t W + ( τ ) B , s γ φ 2 ( τ ) d τ ) 1 γ ω 1 1 γ ( t ) , A 3 = 0 t k k 2 ( 2 k σ α Θ k ) 2 γ 1 ( 2 k σ α Θ k ) ( 2 k ( σ + 1 ) α Θ k ) 2 2 γ 2 ( k k ) ( 1 σ α ) W + B , s d τ 0 t ( sup k 1 2 k σ α Θ k ) 2 γ ( sup k 1 2 k α ( σ + 1 ) Θ k ) 2 2 γ W + B , s d τ ( 0 t W + ( τ ) B , s γ φ 2 ( τ ) d τ ) 1 γ ω 1 1 γ ( t ) ,

and

A 4 = 0 t k k 2 ( 2 k σ α Θ k ) ( 2 k σ α Θ k ) 2 γ 1 ( 2 k ( σ + 1 ) α Θ k ) 2 2 γ 2 ( k k ) ( 1 + σ α + 2 α γ 2 α ) W + B , s d τ 0 t ( sup k 1 2 k σ α Θ k ) 2 γ ( sup k 1 2 k α ( σ + 1 ) Θ k ) 2 2 γ W + B , s d τ ( 0 t W + ( τ ) B , s γ φ 2 ( τ ) d τ ) 1 γ ω 1 1 γ ( t ) .

Combining the above estimates for A 1 - A 4 with (16) and the Young inequality, we have

φ 2 ( t ) φ 2 ( 0 ) + ω ( t ) C ( 0 t W + ( τ ) B , s γ φ 2 ( τ ) d τ ) 1 γ ω 1 1 γ ( t ) C 0 t W + ( τ ) B , s γ φ 2 ( τ ) d τ + ω ( t ) ,

which implies that

φ 2 (t) φ 2 (0)+C 0 t W + ( τ ) B , s γ φ 2 (τ)dτ.

Then the Gronwall inequality gives

φ 2 (t)C φ 2 (0)exp ( C 0 t W + ( τ ) B , s γ d τ ) .
(17)

Case II: s(1,0). Let

φ(t) sup k 1 2 k ( 1 1 γ ) Θ k (t),ω(t) sup k 1 2 2 k α γ 0 t Θ k 2 γ (τ)dτ.

Multiplying (15) by 2 2 k α ( γ 1 ) Θ k 2 ( γ 1 ) , and integrating with respect to t, it follows that

2 2 k α ( γ 1 ) Θ k 2 γ ( t ) 2 2 k α ( γ 1 ) Θ k 2 γ ( 0 ) + 2 2 k α γ 0 t Θ k 2 γ ( τ ) d τ 0 t k k 2 2 2 k α ( γ 1 ) Θ k 2 ( γ 1 ) 2 2 k α ( 1 1 γ ) Θ k 2 W + B , s d τ + 0 t k k 2 2 2 k α ( γ 1 ) Θ k 2 ( γ 1 ) 2 2 k α ( 1 1 γ ) Θ k 2 W + B , s d τ + 0 t k k 2 2 2 k α ( γ 1 ) Θ k 2 ( γ 1 ) 2 k 2 k ( 2 α 1 2 α γ ) Θ k Θ k W + B , s d τ + 0 t k k 2 2 2 k α ( γ 1 ) Θ k 2 ( γ 1 ) 2 k 2 k ( 2 α 1 2 α γ ) Θ k Θ k W + B , s d τ B 1 + B 2 + B 3 + B 4 .
(18)

By Lemma 2.2 and the Hölder inequality, we have

B 1 = 0 t k k 2 ( 2 2 k α γ Θ k 2 γ ) 1 1 γ ( 2 k α ( 1 1 γ ) Θ k ) 2 2 2 α ( k k ) ( 1 1 γ ) W + B , s d τ 0 t ( sup k 1 2 2 k α γ Θ k 2 γ ) 1 1 γ ( sup k 1 2 k α ( 1 1 γ ) Θ k ) 2 W + B , s d τ ( 0 t W + ( τ ) B , s γ φ 2 γ ( τ ) d τ ) 1 γ ω 1 1 γ ( t ) , B 2 = 0 t k k 2 ( 2 2 k α γ Θ k 2 γ ) 1 1 γ ( 2 k α ( 1 1 γ ) Θ k ) 2 2 2 α ( k k ) ( 1 1 γ ) W + B , s d τ 0 t ( sup k 1 2 2 k α γ Θ k 2 γ ) 1 1 γ ( sup k 1 2 k α ( 1 1 γ ) Θ k ) 2 W + B , s d τ ( 0 t W + ( τ ) B , s γ φ 2 γ ( τ ) d τ ) 1 γ ω 1 1 γ ( t ) , B 3 = 0 t k k 2 ( 2 2 k α γ Θ k 2 γ ) 1 1 γ 2 k α ( 1 1 γ ) Θ k 2 k α ( 1 1 γ ) Θ k B 3 = 2 ( k k ) ( 1 α + α γ ) W + B , s d τ B 3 0 t ( sup k 1 2 2 k α γ Θ k 2 γ ) 1 1 γ ( sup k 1 2 k α ( 1 1 γ ) Θ k ) 2 W + B , s d τ B 3 ( 0 t W + ( τ ) B , s γ φ 2 γ ( τ ) d τ ) 1 γ ω 1 1 γ ( t ) ,

and

B 4 = 0 t k k 2 ( 2 2 k α γ Θ k 2 γ ) 1 1 γ 2 k α ( 1 1 γ ) Θ k 2 k α ( 1 1 γ ) Θ k 2 ( k k ) ( 1 α + α γ ) W + B , s d τ 0 t ( sup k 1 2 2 k α γ Θ k 2 γ ) 1 1 γ ( sup k 1 2 k α ( 1 1 γ ) Θ k ) 2 W + B , s d τ ( 0 t W + ( τ ) B , s γ φ 2 γ ( τ ) d τ ) 1 γ ω 1 1 γ ( t ) .

Inserting the above estimates for B 1 - B 4 into (18) and by the Young inequality, we have

φ 2 γ ( t ) φ 2 γ ( 0 ) + ω ( t ) C ( 0 t W + ( τ ) B , s γ φ 2 γ ( τ ) d τ ) 1 γ ω 1 1 γ ( t ) C 0 t W + ( τ ) B , s γ φ 2 γ ( τ ) d τ + ω ( t ) ,

which gives

φ 2 γ (t) φ 2 γ (0)+C 0 t W + ( τ ) B , s γ φ 2 γ (τ)dτ.

Again, by the Gronwall inequality, we have

sup t ( 0 , T ] φ 2 γ (t) φ 2 γ (0)exp ( C 0 t W + ( τ ) B , s γ d τ ) .
(19)

By Lemma 2.3, as s[0,1), for some η<σα, the following embedding holds:

B 2 , σ α ( R 3 ) B 2 , 2 η ( R 3 ) = H η ( R 3 ) .

According to the local existence with the condition [5]η> 5 2 2α, thus σ can be taken as 5 2 2 α α <σ< 1 α . When s(1,0), for α(1, 5 4 ), one has 1 s 2 5 2 2α, which infers that

B 2 , 1 s 2 ( R 3 ) B 2 , 2 η ( R 3 ) = H η ( R 3 ) .

Then for any qγ, if W + L q (0,T; B , s ( R 3 )) with 2 α q 2 α γ =2α1+s, by the standard Picard method [15], [16], we can easily show that the solution (u,b) remains smooth at time t=T.

References

  1. 1.

    Wu J: Generalized MHD equations. J. Differ. Equ. 2003, 195: 284-312. 10.1016/j.jde.2003.07.007

    Article  Google Scholar 

  2. 2.

    Wu J: Regularity criteria for the generalized MHD equations. Commun. Partial Differ. Equ. 2008, 33: 285-360. 10.1080/03605300701382530

    Article  Google Scholar 

  3. 3.

    Zhou Y: Regularity criteria for the generalized viscous MHD equations. Ann. Inst. Henri Poincaré, Anal. Non Linéaire 2007, 24: 491-505. 10.1016/j.anihpc.2006.03.014

    Article  Google Scholar 

  4. 4.

    Wu G: Regularity criteria for the 3D generalized MHD equations in terms of vorticity. Nonlinear Anal. 2009, 71: 4251-4258. 10.1016/j.na.2009.02.115

    MathSciNet  Article  Google Scholar 

  5. 5.

    Yuan J: Existence theorem and regularity criteria for the generalized MHD equations. Nonlinear Anal., Real World Appl. 2010, 11: 1640-1649. 10.1016/j.nonrwa.2009.03.017

    MathSciNet  Article  Google Scholar 

  6. 6.

    Hasegawa A: Self-organization processes in continuous media. Adv. Phys. 1985, 34: 1-42. 10.1080/00018738500101721

    Article  Google Scholar 

  7. 7.

    Politano H, Pouquet A, Sulem PL: Current and vorticity dynamics in three-dimensional magnetohydrodynamics turbulence. Phys. Plasmas 1995, 2: 2931-2939. 10.1063/1.871473

    MathSciNet  Article  Google Scholar 

  8. 8.

    He C, Wang Y: Remark on the regularity for weak solutions to the magnetohydrodynamic equations. Math. Methods Appl. Sci. 2008, 31: 1667-1684. 10.1002/mma.992

    MathSciNet  Article  Google Scholar 

  9. 9.

    Gala S: Extension criterion on regularity for weak solutions to the 3D MHD equations. Math. Methods Appl. Sci. 2010, 33: 1496-1503.

    MathSciNet  Google Scholar 

  10. 10.

    Dong B, Jia Y, Zhang W: An improved regularity criterion of three-dimensional magnetohydrodynamic equations. Nonlinear Anal., Real World Appl. 2012, 13: 1159-1169. 10.1016/j.nonrwa.2011.09.010

    MathSciNet  Article  Google Scholar 

  11. 11.

    Cannone M: Harmonic analysis tools for solving the incompressible Navier-Stokes equations. Handbook of Mathematical Fluid Dynamics 2005, 161-244. 10.1016/S1874-5792(05)80006-0

    Chapter  Google Scholar 

  12. 12.

    Bahouri H, Chemin JY, Danchin R: Fourier Analysis and Nonlinear Partial Differential Equations. Springer, Berlin; 2011.

    Book  Google Scholar 

  13. 13.

    Chemin JY: Perfect Incompressible Fluids. Oxford University Press, New York; 1998.

    Google Scholar 

  14. 14.

    Brenner P: Besov Spaces and Applications to Difference Methods for Initial Value Problems. Springer, Berlin; 1975.

    Google Scholar 

  15. 15.

    Fujita H, Kato T: On the Navier-Stokes initial value problem I. Arch. Ration. Mech. Anal. 1964, 16: 269-315. 10.1007/BF00276188

    MathSciNet  Article  Google Scholar 

  16. 16.

    Kato T:Nonstationary flows of viscous and ideal fluids in R 3 . J. Funct. Anal. 1972, 9: 296-305. 10.1016/0022-1236(72)90003-1

    Article  Google Scholar 

Download references

Acknowledgements

The authors are highly grateful for the referees’ careful reading and comments on this paper. The work of Zhang is partially supported by the NNSF of China under the grant 11101337, Doctoral Foundation of Ministry of Education of China grant-20110182120013, Fundamental Research Funds for the Central Universities grant-XDJK2011C046, and Doctoral Fund of Southwest University (SWU110035). The work of Qiu is partially supported by the NNSF of China grant-11126266, the NSF of Guangdong grant-S2013010013608, Foundation for Distinguished Young Talents in Higher Education of Guangdong, China grant 2012LYM_0030 and Pearl River New Star Program grant-2012J2200016.

Author information

Affiliations

Authors

Corresponding author

Correspondence to Hua Qiu.

Additional information

Competing interests

The authors declare that they have no competing interests.

Authors’ contributions

All authors contributed equally to the writing of this paper. All authors read and approved the final manuscript.

Rights and permissions

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0), which permits use, duplication, 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 license, and indicate if changes were made.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Zhang, S., Qiu, H. On the regularity criterion for the 3D generalized MHD equations in Besov spaces. Bound Value Probl 2014, 178 (2014). https://doi.org/10.1186/s13661-014-0178-3

Download citation

Keywords

  • generalized MHD equations
  • regularity criterion
  • Littlewood-Paley decomposition