A Liouville-type theorem for the stationary MHD equations (Q6100977)

The goal of this paper is to establish a Liouville-type theorem for the stationary MHD equations in $R^3$. The equations under consideration are \begin{gather*} -\Delta \mathbf{u} + \mathbf{u}\cdot \nabla \mathbf{u} - \mathbf{B}\cdot \nabla \mathbf{B} + \nabla \pi = \mathbf{0}, \tag{1}\\ -\Delta \mathbf{B} + \mathbf{u} \cdot \nabla \mathbf{B} - \mathbf{B} \cdot \nabla \mathbf{u} = 0, \tag{2}\\ \nabla \cdot \mathbf{u} = \nabla \cdot \mathbf{B} = 0 \tag{3} \end{gather*} Here $\mathbf{u}$ is the velocity field, $\mathbf{B}$ is the magnetic field, and $\pi = p + \frac{1}{2}|\mathbf{B}|^2$, where $p$ is the pressure in the fluid. The kinematic viscosity coefficient $\nu$ in front of $\nabla \mathbf{u}$ and the coefficient of magnetic diffusion $\xi$ in front of $\nabla \mathbf{B}$ in the MHD equations are both set to $\nu = \xi = 1$ as they do not play a role in the analysis in the paper. A Liouville-type theorem is a theorem that proves that under certain fixed growth conditions on the solution the only solution to the equations under consideration is the trivial solution. Such theorems have previously been proved for the stationary Navier-Stokes equations. The objective of the present paper is to extend this type of analysis to the equations (1), (2), (3) above. Previously Louiville type theorems have been proven for the stationary MHD equations as cited in the paper. The authors go beyond this prior work in the following way. Studies have shown that the velocity field $\mathbf{u}$ plays a more dominant role in the solutions of the MHD equations than the magnetic field $\mathbf{B}$. Following this the authors impose a fixed condition on the mean oscillation of the velocity field but then allow more freedom in applying conditions to the magnetic field than has been done in previous papers on the subject. The Liouville-type theorem for equations (1), (2), (3) proved in the paper relies on the following assumptions on $\mathbf{u}$ and $\mathbf{B}$. It is assumed that $(\mathbf{u}, \mathbf{B})$ is a weak solution to (1), (2), (3), and that $(\mathbf{u}, \mathbf{B})$ can be represented as $\mathbf{u} = \nabla \cdot \Phi$ and $\mathbf{B} = \nabla \cdot \Psi$ for $\Phi,\Psi \in C^{\infty}(R^3,R^{3 \times 3})$. In addition the tensor valued functions $\Phi,\Psi$ satisfy the estimates \begin{itemize} \item[a)] \(\sup R^{-\beta}_{R>1} \Big (\int_{B_R(\mathbf{0})} |\Phi - \Phi_{B_R(\mathbf{0}}|^{\alpha} d\mathbf{x} \Big )^ {\frac{1}{\alpha}} < \infty\), \item[b)] \(\sup R^{-\delta}_{R>1} \Big (\int_{B_R(\mathbf{0})} |\Psi - \Psi_{B_R(\mathbf{0}}|^{\gamma} d\mathbf{x} \Big )^ {\frac{1}{\gamma}} < \infty\), \end{itemize} in which the $\alpha,\beta,\delta,\gamma$ satisfy \begin{itemize} \item[a)] \(3 < \alpha \le \frac{8+2\sqrt{7}}{3}\), \(0 \leq \beta \leq \frac{1}{3} - \frac{1}{\alpha}\), \item[b)] \(\frac{12\alpha-18}{5\alpha-9} < \gamma \leq \frac{10\alpha-12}{3\alpha-6}\), \(0 \leq \delta \leq \frac{(\gamma+6)(5\alpha-9-\alpha \beta}{2\gamma(7\alpha-12)}-\frac{3}{\gamma}\). \end{itemize} The main result of this paper is then given as a theorem that proves that under the above stated assumptions the solution to the equations (1), (2), (3) is $\mathbf{u} = \mathbf{B} = \mathbf{0}$. The authors point out that for specific values of $\alpha,\beta,\gamma,\delta$ The present theorem yields the results obtained in references [5], [6] cited in the paper.