This article studies a class of variational inequality problems composed of non-divergence type parabolic operators. In comparison with traditional differential equations, this study focuses on overcoming inequality constraints to obtain Hölder and Schauder estimates for weak solutions. The results indicate that the weak solution of the variational inequality possesses the Cα continuity and the Schauder estimate on the W1,p space, where α∈(0,1) and p≥2.
Citation: Yudong Sun, Tao Wu. Hölder and Schauder estimates for weak solutions of a certain class of non-divergent variation inequality problems in finance[J]. AIMS Mathematics, 2023, 8(8): 18995-19003. doi: 10.3934/math.2023968
[1] | Xiao Guo, Chuanpei Xu, Zhibin Zhu, Benxin Zhang . Nonmonotone variable metric Barzilai-Borwein method for composite minimization problem. AIMS Mathematics, 2024, 9(6): 16335-16353. doi: 10.3934/math.2024791 |
[2] | Jamilu Sabi'u, Ali Althobaiti, Saad Althobaiti, Soubhagya Kumar Sahoo, Thongchai Botmart . A scaled Polak-Ribiˊere-Polyak conjugate gradient algorithm for constrained nonlinear systems and motion control. AIMS Mathematics, 2023, 8(2): 4843-4861. doi: 10.3934/math.2023241 |
[3] | Jamilu Sabi'u, Ibrahim Mohammed Sulaiman, P. Kaelo, Maulana Malik, Saadi Ahmad Kamaruddin . An optimal choice Dai-Liao conjugate gradient algorithm for unconstrained optimization and portfolio selection. AIMS Mathematics, 2024, 9(1): 642-664. doi: 10.3934/math.2024034 |
[4] | Sani Aji, Aliyu Muhammed Awwal, Ahmadu Bappah Muhammadu, Chainarong Khunpanuk, Nuttapol Pakkaranang, Bancha Panyanak . A new spectral method with inertial technique for solving system of nonlinear monotone equations and applications. AIMS Mathematics, 2023, 8(2): 4442-4466. doi: 10.3934/math.2023221 |
[5] | Yiting Zhang, Chongyang He, Wanting Yuan, Mingyuan Cao . A novel nonmonotone trust region method based on the Metropolis criterion for solving unconstrained optimization. AIMS Mathematics, 2024, 9(11): 31790-31805. doi: 10.3934/math.20241528 |
[6] | Ting Lin, Hong Zhang, Chaofan Xie . A modulus-based modified multivariate spectral gradient projection method for solving the horizontal linear complementarity problem. AIMS Mathematics, 2025, 10(2): 3251-3268. doi: 10.3934/math.2025151 |
[7] | Zhensheng Yu, Peixin Li . An active set quasi-Newton method with projection step for monotone nonlinear equations. AIMS Mathematics, 2021, 6(4): 3606-3623. doi: 10.3934/math.2021215 |
[8] | Luyao Zhao, Jingyong Tang . Convergence properties of a family of inexact Levenberg-Marquardt methods. AIMS Mathematics, 2023, 8(8): 18649-18664. doi: 10.3934/math.2023950 |
[9] | Austine Efut Ofem, Jacob Ashiwere Abuchu, Godwin Chidi Ugwunnadi, Hossam A. Nabwey, Abubakar Adamu, Ojen Kumar Narain . Double inertial steps extragadient-type methods for solving optimal control and image restoration problems. AIMS Mathematics, 2024, 9(5): 12870-12905. doi: 10.3934/math.2024629 |
[10] | Limei Xue, Jianmin Song, Shenghua Wang . A modified projection and contraction method for solving a variational inequality problem in Hilbert spaces. AIMS Mathematics, 2025, 10(3): 6128-6143. doi: 10.3934/math.2025279 |
This article studies a class of variational inequality problems composed of non-divergence type parabolic operators. In comparison with traditional differential equations, this study focuses on overcoming inequality constraints to obtain Hölder and Schauder estimates for weak solutions. The results indicate that the weak solution of the variational inequality possesses the Cα continuity and the Schauder estimate on the W1,p space, where α∈(0,1) and p≥2.
In this paper, we consider the two-dimensional viscous, compressible and heat conducting magnetohydrodynamic equations in the Eulerian coordinates (see [1])
{ρt+div(ρu)=0,(ρu)t+div(ρu⊗u)+∇P=μ△u+(μ+λ)∇(divu)+H⋅∇H−12∇|H|2,cv((ρθ)t+div(ρuθ))+Pdivu=κΔθ+λ(divu)2+ν|curlH|2+2μ|D(u)|2,Ht+u⋅∇H−H⋅∇u+Hdivu=νΔH,divH=0. | (1.1) |
Here x=(x1,x2)∈Ω is the spatial coordinate, Ω is a bounded smooth domain in R2, t≥0 is the time, and the unknown functions ρ=ρ(x,t), θ=θ(x,t), u=(u1,u2)(x,t) and H=(H1,H2)(x,t) denote, respectively, the fluid density, absolute temperature, velocity and magnetic field. In addition, the pressure P is given by
P(ρ)=Rθρ,(R>0), |
where R is a generic gas constant. The deformation tensor D(u) is defined by
D(u)=12(∇u+(∇u)tr). |
The shear viscosity μ and the bulk one λ satisfy the hypotheses as follows
μ>0,μ+λ≥0. |
Positive constants cv, κ and ν represent, respectively, the heat capacity, heat conductivity and magnetic diffusivity coefficient.
The initial condition and boundary conditions for Eq (1.1) are given as follows
(ρ,θ,u,H)(x,t=0)=(ρ0,θ0,u0,H0), | (1.2) |
∂θ∂n=0,u=0,H⋅n=0,curlH=0,on∂Ω, | (1.3) |
where n denotes the unit outward normal vector of ∂Ω.
Remark 1.1. The boundary condition imposed on H (1.3) is physical and means that the container is perfectly conducting, see [1,2,3,4].
In the absence of electromagnetic effect, namely, in the case of H≡0, the MHD system reduces to the Navier-Stokes equations. Due to the strong coupling and interplay interaction between the fluid motion and the magnetic field, it is rather complicated to investigate the well-posedness and dynamical behaviors of MHD system. There are a huge amount of literature on the existence and large time behavior of solutions to the Navier-Stokes system and MHD one due to the physical importance, complexity, rich phenomena and mathematical challenges, see [1,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26] and the reference therein. However, many physically important and mathematically fundamental problems are still open due to the lack of smoothing mechanism and the strong nonlinearity. When the initial density contain vacuum states, the local large strong solutions to Cauchy problem of 3D full MHD equations and 2D isentropic MHD system have been obtained, respectively, by Fan-Yu [5] and Lü-Huang [6]. For the global well-posedness of strong solutions, Li-Xu-Zhang [7] and Lü-Shi-Xu [8] established the global existence and uniqueness of strong solutions to the 3D and 2D MHD equations, respectively, provided the smooth initial data are of small total energy. In particular, the initial density can have compact support in [7,8]. Furthermore, Hu-Wang [9,10] and Fan-Yu [11] proved the global existence of renormalized solutions to the compressible MHD equations for general large initial data. However, it is an outstanding challenging open problem to establish the global well-posedness for general large strong solutions with vacuum.
Therefore, it is important to study the mechanism of blow-up and structure of possible singularities of strong (or smooth) solutions to the compressible MHD system (1.1). The pioneering work can be traced to Serrin's criterion [12] on the Leray-Hopf weak solutions to the 3D incompressible Navier-Stokes equations, that is
limt→T∗‖u‖Ls(0,t;Lr)=∞,for3r+2s=1,3<r≤∞, | (1.4) |
where T∗ is the finite blow up time. Later, He-Xin [13] established the same Serrin's criterion (1.4) for strong solutions to the incompressible MHD equations.
First of all, we recall several known blow-up criteria for the compressible Navier-Stokes equations. In the isentropic case, Huang-Li-Xin [14] established the Serrin type criterion as follows
limt→T∗(‖u‖Ls(0,t;Lr)+‖divu‖L1(0,t;L∞))=∞,for3r+2s=1,3<r≤∞. | (1.5) |
For the full compressible Navier-Stokes equations, Fan-Jiang-Ou [15] obtained that
limt→T∗(‖θ‖L∞(0,t;L∞)+‖∇u‖L1(0,t;L∞))=∞, | (1.6) |
under the condition
7μ>λ. | (1.7) |
Later, the restriction (1.7) was removed in Huang-Li-Xin [16]. Recently, Wang [17] established a blow-up criterion for the initial boundary value problem (IBVP) on a smooth bounded domain in R2, namely,
limt→T∗‖divu‖L1(0,t;L∞)=∞. | (1.8) |
Then, let's return to the compressible MHD system (1.1). Under the three-dimensional isentropic condition, Xu-Zhang [18] founded the same criterion (1.5) as [14]. For the three-dimensional full compressible MHD system, the criterion (1.6) is also established by Lu-Du-Yao [19] under the condition
μ>4λ. | (1.9) |
Soon, the restriction (1.9) was removed by Chen-Liu [20]. Later, for the Cauchy problem and the IBVP of three-dimensional full compressible MHD system, Huang-Li [21] proved that
limt→T∗(‖u‖Ls(0,t;Lr)+‖ρ‖L∞(0,t;L∞))=∞,for3r+2s≤1,3<r≤∞. | (1.10) |
Recently, Fan-Li-Nakamura [22] extended the results of [17] to the MHD system and established a blow-up criterion which depend only on H and divu as follows
limt→T∗(‖H‖L∞(0,t;L∞)+‖divu‖L1(0,t;L∞))=∞. | (1.11) |
In fact, if H≡0 in (1.11), the criterion (1.11) becomes (1.8).
The purpose of this paper is to loosen and weaken the regularity of H required in the blow-up criterion (1.11) for strong solutions of the IBVP (1.1)–(1.3).
In this paper, we denote
∫⋅dx≜ |
Furthermore, for s\ge 0 and 1\le r\le \infty , we define the standard Lebesgue and Sobolev spaces as follows
\begin{equation} \notag\begin{cases} L^r = L^r( \Omega), \quad W^{s, r} = W^{s, r}( \Omega), \quad H^s = W^{s, 2}, \\ W^{s, r}_0 = \{f\in W^{s, r}|f = 0\; \mbox{on}\; {\partial} \Omega\}, \quad H_0^s = W_0^{s, 2}. \end{cases} \end{equation} |
To present our results, we first recall the local existence theorem of the strong solution. Fan-Yu [5] attained the local existence and uniqueness of strong solution with full compressible MHD system in \mathbb{R}^3 . In fact, when \Omega is a bounded domain in \mathbb{R}^2 , the method applied in [5,23] can also be used to the case here. The corresponding result can be expressed as follows.
Theorem 1.1. (Local existence theorem) For q > 2 , assume that the initial data (\rho_0, \theta_0, u_0, H_0) satisfies
\begin{equation} \begin{cases} 0\le\rho_0\in W^{1, q}, \, \, \; \; 0\le \theta_0\in H^2, \; \; u_0 \in H_0^1\cap H^2, \; \; H_0 \in H^2, \, \, \; \; \; {{\rm{div }}} H_0 = 0, \\ \frac{{\partial} \theta_0}{{\partial} n}|_{{\partial} \Omega} = 0, \; \; \; u_0|_{{\partial} \Omega} = 0, \; \; \; H_0\cdot n|_{{\partial} \Omega} = \mathit{\mbox{curl}} H_0|_{{\partial} \Omega} = 0, \end{cases} \end{equation} | (1.12) |
and the compatibility conditions as follows
\begin{equation} - \mu{\triangle} u_0 - (\mu + {\lambda})\nabla {{\rm{div }}} u_0 + R\nabla ( \rho_0 \theta_0)-H_0\cdot \nabla H_0+\frac{1}{2} \nabla |H_0|^2 = \rho_0^{1/2}g_1, \end{equation} | (1.13) |
\begin{equation} - \kappa{\triangle} \theta_0 - 2\mu|\mathfrak{D}(u_0)|^2-\lambda( {{\rm{div }}} u_0)^2-\nu( \mathit{\mbox{curl}} H_0)^2 = \rho_0^{1/2}g_2, \end{equation} | (1.14) |
for some g_1, \; g_2\in L^2 . Then there exists a time T_0 > 0 such that the IBVP (1.1)–(1.3) has a unique strong solution (\rho, \theta, u, H) on \Omega\times (0, T_0] satisfying that
\begin{equation} \begin{cases} 0\le \rho\in C([0, T_0]; W^{1, q} ), \; \; \; \rho_t\in C([0, T_0]; L^q ), \\ (u, \theta, H)\in C([0, T_0]; H^2)\cap L^2(0, T_0; W^{2, q}), \; \; \; \theta\ge 0, \\ (u_t, \theta_t, H_t)\in L^2(0, T_0; H^1), \; \; \; (\sqrt{ \rho} u_t, \sqrt{ \rho} \theta_t, H_t)\in L^\infty(0, T_0; L^2). \end{cases} \end{equation} | (1.15) |
Then, our main result is stated as follows.
Theorem 1.2. Under the assumption of Theorem 1.1, suppose (\rho, \theta, u, H) is the strong solution of the IBVP (1.1)–(1.3) obtained in Theorem 1.1. If T^* < \infty is the maximum existence time of the strong solution, then
\begin{equation} \lim\limits_{t\rightarrow T^*} \left(\|H\|_{L^\infty(0, t; L^b)}+\| {{\rm{div }}} u\|_{L^1(0, t; L^\infty)} \right) = \infty, \end{equation} | (1.16) |
for any b > 2 .
Remark 1.2. Compared to the blow-up criterion (1.11) attained in [22], Theorem 1.2 demonstrates some new message about the blow-up mechanism of the MHD system (1.1)–(1.3). Particularly, beside the same regularity on \| {{\rm{div }}} u\|_{L^1(0, t; L^\infty)} as (1.11) in [22], our result (1.16) improves the regularity on \|H\|_{L^\infty(0, t; L^\infty)} by relaxing it to \|H\|_{L^\infty(0, t; L^b)} for any b > 2 .
The rest of the paper is arranged as follows. We state several basic facts and key inequalities which are helpful for later analysis in Section 2. Sections 3 is devoted to a priori estimate which is required to prove Theorem 1.2, while we give its proof in Section 4.
In this section, we will recall several important inequalities and well-known facts. First of all, Gagliardo-Nirenberg inequality (see [27]) is described as follows.
Lemma 2.1. (Gagliardo-Nirenberg) For q\in(1, \infty), r\in (2, \infty) and s\in [2, \infty), there exists some generic constant C > 0 which may depend only on q, r and s such that for f\in C_0^\infty(\Omega) , we have
\begin{equation} \|f\|_{L^s(\Omega)}^s\le C \|f\|_{L^2(\Omega)}^{2}\| \nabla f\|_{L^2(\Omega)}^{s-2} , \end{equation} | (2.1) |
\begin{equation} \|g\|_{L^\infty(\Omega)} \le C \|g\|_{L^q(\Omega)}^{q(r-2)/(2r+q(r-2))}\| \nabla g\|_{L^r(\Omega)}^{2r/(2r+q(r-2))} . \end{equation} | (2.2) |
Then, we give several regularity results for the following Lamé system with Dirichlet boundary condition (see [24])
\begin{equation} \begin{aligned}\begin{cases} \mathcal{L}U\triangleq \mu\Delta U+(\mu+\lambda) \nabla {{\rm{div }}} U = F, \, \, &x\in\Omega, \\ U = 0, \, \, &x\in \partial\Omega.\end{cases} \end{aligned} \end{equation} | (2.3) |
We assume that U\in H^1_0 is a weak solution of the Lamé system, due to the uniqueness of weak solution, it could be denoted by U = \mathcal{L}^{-1}F .
Lemma 2.2. Let r\in (1, \infty), then there exists some generic constant C > 0 depending only on \mu, \lambda, r and \Omega such that
● If F\in L^r , then
\begin{equation} \|U\|_{W^{2, r}(\Omega)} \le C \|F\|_{L^r(\Omega)}. \end{equation} | (2.4) |
● If F\in W^{-1, r} (i.e., F = {{\rm{div }}} f with f = (f_{ij})_{2\times2}, f_{ij}\in L^r ), then
\begin{equation} \|U\|_{W^{1, r}(\Omega)} \le C \|f\|_{L^r(\Omega)}. \end{equation} | (2.5) |
Furthermore, for the endpoint case, if f_{ij}\in L^2\cap L^\infty , then \nabla U\in \mathit{\mbox{BMO}}(\Omega) and
\begin{equation} \| \nabla U\|_{\mathit{\mbox{BMO}}(\Omega)} \le C \|f\|_{L^\infty(\Omega)}+ C \|f\|_{L^2(\Omega)}. \end{equation} | (2.6) |
The following L^p -bound for elliptic systems, whose proof is similar to that of [28,Lemma 12], is a direct consequence of the combination of a well-known elliptic theory due to Agmon-Douglis-Nirenberg[29,30] with a standard scaling procedure.
Lemma 2.3. For k\ge 0 and p > 1 , there exists a constant C > 0 depending only on k and p such that
\begin{equation} \| \nabla^{k+2}v\|_{L^p( \Omega)}\le C\| \Delta v\|_{W^{k, p}( \Omega)}, \end{equation} | (2.7) |
for every v\in W^{k+2, p}(\Omega) satisfying either
\begin{eqnarray*} v\cdot n = 0, \, \, {{\rm{rot}}} v = 0, \, \, \mathit{\mbox{on}} ~~ \partial \Omega, \end{eqnarray*} |
or
\begin{eqnarray*} v = 0, \, \, \mathit{\mbox{on}}~~ \partial \Omega. \end{eqnarray*} |
Finally, we give two critical Sobolev inequalities of logarithmic type, which are originally due to Brezis-Gallouet [31] and Brezis-Wainger [32].
Lemma 2.4. Let \Omega\subset{\mathbb{R}^2} be a bounded Lipschitz domain and f\in W^{1, q} with q > 2 , then it holds that
\begin{equation} \|f\|_{L^\infty( \Omega)} \le C \|f\|_{\mathit{\mbox{BMO}}( \Omega)}\ln\left(e+\|f\|_{W^{1, q}( \Omega)}\right)+C, \end{equation} | (2.8) |
with a constant C depending only on q .
Lemma 2.5. Let \Omega\subset{\mathbb{R}^2} be a smooth domain and f\in L^2(s, t; H^1_0\cap W^{1, q}) with q > 2 , then it holds that
\begin{equation} \|f\|_{L^2(s, t; L^\infty)}^2\le C \|f\|_{L^2(s, t; H^1)}^2 \ln\left(e+\|f\|_{L^2(s, t; W^{1, q})}\right)+C, \end{equation} | (2.9) |
with a constant C depending only on q .
Let (\rho, \theta, u, H) be the strong solution of the IBVP (1.1)–(1.3) obtained in Theorem 1.1. Assume that (1.16) is false, namely, there exists a constant M > 0 such that
\begin{equation} \lim\limits_{t\rightarrow T^*} \left(\| H\|_{L^\infty(0, t; L^b)}+\| {{\rm{div }}} u\|_{L^1(0, t; L^\infty)} \right)\le M < \infty, \; \; \; \; \mbox{for any}\; b > 2. \end{equation} | (3.1) |
First of all, the upper bound of the density can be deduced from (1.1) _1 and (3.1), see [14,Lemma 3.4].
Lemma 3.1. Under the assumptions of Theorem 1.2 and (3.1), it holds that for any t\in [0, T^*),
\begin{equation} \begin{aligned} \sup\limits_{0\le s\le t}\| \rho\|_{L^1\cap L^\infty}\le C, \end{aligned} \end{equation} | (3.2) |
where (and in what follows) C represents a generic positive constant depending only on \mu, \lambda, c_v, \kappa , \nu , q, b, M , T^* and the initial data.
Then, we give the following estimates, which are similar to the energy estimates.
Lemma 3.2. Under the assumptions of Theorem 1.2 and (3.1), it holds that for any t\in [0, T^*),
\begin{equation} \begin{aligned} \sup\limits_{0\le s\le t} \left(\| \rho \theta\|_{L^1}+\|\rho^{1/2}u\|^2_{L^{2}}+\| H\|_{L^2}^2\right) +\int_{0}^{t}\left(\| \nabla u\|_{L^2}^2+\| \nabla H\|_{L^2}^2 \right)ds \le C. \end{aligned} \end{equation} | (3.3) |
Proof. First, using the standard maximum principle to (1.1) _3 together with \theta_0\ge0 (see [15,25]) gives
\begin{equation} \begin{aligned} \inf\limits_{ \Omega\times[0, t]} \theta(x, t)\ge 0. \end{aligned} \end{equation} | (3.4) |
Then, utilizing the standard energy estimates to (1.1) shows
\begin{equation} \begin{aligned} \sup\limits_{0\le s\le t} \left(\| \rho \theta\|_{L^1}+\|\rho^{1/2}u\|^2_{L^{2}}+\| H\|_{L^2}^2\right)\le C. \end{aligned} \end{equation} | (3.5) |
Next, adding (1.1)_{2} multiplied by u to (1.1)_{4} multiplied by H , and integrating the summation by parts, we have
\begin{equation} \begin{aligned} &\frac{1}{2}\frac{d}{dt}\left(\|\rho^{1/2}u\|^2_{L^{2}} +\| H\|_{L^2}^2\right)\\ &\quad+\mu\| \nabla u\|_{L^2}^2+ \nu \| \nabla H\|_{L^2}^2+(\mu+\lambda)\| {{\rm{div }}} u\|_{L^2}^2\le C\| \rho \theta\|_{L^1}\| {{\rm{div }}} u\|_{L^\infty}, \end{aligned} \end{equation} | (3.6) |
where one has used the following well-known fact
\begin{equation} \begin{aligned} \| \nabla H\|_{L^2} \le C\| \mbox{curl} H\|_{L^2}, \end{aligned} \end{equation} | (3.7) |
due to {{\rm{div }}} H = 0 and H\cdot n|_{{\partial} \Omega} = 0 .
Hence, the combination of (3.6) with (3.1), (3.4) and (3.5) yields (3.3). This completes the proof of Lemma 3.2.
The following lemma shows the estimates on the spatial gradients of both the velocity and the magnetic, which are crucial for obtaining the higher order estimates of the solution.
Lemma 3.3. Under the assumptions of Theorem 1.2 and (3.1), it holds that for any t\in [0, T^*),
\begin{equation} \begin{aligned} &\sup\limits_{0\le s\le t} \left( \| \rho^{1/2} \theta\|_{L^2}^2+ \| \nabla u\|_{L^2}^2+ \| \mathit{\mbox{curl}} H\|_{L^2}^2\right)\\ &\quad+\int_{0}^{t} \left( \| \rho^{1/2}\dot u\|_{L^2}^2+\| \nabla \theta\|_{L^2}^2 + \|H_t\|_{L^2}^2+ \| \Delta H\|_{L^2}^2 \right)ds \le C, \end{aligned} \end{equation} | (3.8) |
where \dot f\triangleq u\cdot \nabla f+f_t represents the material derivative of f .
Proof. Above all, multiplying the equation (1.1)_{3} by \theta and integrating by parts yield
\begin{equation} \begin{aligned} \frac{c_v}{2} \frac{d}{dt} \|\rho^{1/2} \theta\|^2_{L^{2}} +\kappa\| \nabla \theta\|_{L^2}^2 &\le\nu\int \theta|\mbox{curl} H|^2dx +C\int \theta| \nabla u|^2dx+C\| \rho^{1/2} \theta\|_{L^2}^2\| {{\rm{div }}} u\|_{L^\infty}. \end{aligned} \end{equation} | (3.9) |
Firstly, integration by parts together with (3.1) and Gagliardo-Nirenberg inequality implies that
\begin{equation} \begin{aligned} \nu\int \theta| \mbox{curl} H|^2dx\le &C \| \nabla \theta\|_{L^2}\|H\|_{L^b}\| \nabla H\|_{L^{\tilde b}}+C\| \theta\|_{L^{\tilde b}}\|H\|_{L^b} \| \nabla^2H\|_{L^2}\\ \le & C\| \nabla \theta\|_{L^2}\| \nabla H\|_{L^{\tilde b}} +C\| \nabla^2H\|_{L^2}(\| \nabla \theta\|_{L^2}+1) \\ \le & \varepsilon\| \nabla \theta\|_{L^2}^2+C\| \nabla^2 H\|_{L^2}^2 +C(\| \nabla H\|_{L^2}^2+1), \end{aligned} \end{equation} | (3.10) |
where \tilde b\triangleq \frac{2b}{b-2} > 2 satisfies 1/b+1/\tilde b = 1/2 , and in the second inequality where one has applied the estimate as follows
\begin{equation} \begin{aligned} \| \theta\|_{L^r}\le C(\| \nabla \theta\|_{L^2}+1), \; \; \; \; \mbox{for any}\; r\ge 1. \end{aligned} \end{equation} | (3.11) |
Indeed, denote the average of \theta by \bar \theta = \frac{1}{| \Omega|}\int \theta dx , it follows from (3.2) and (3.3) that
\begin{equation} \begin{aligned} \bar \theta\int \rho dx\le \int \rho \theta dx+\int \rho | \theta-\bar \theta|dx\le C+C\| \nabla \theta\|_{L^2}, \end{aligned} \end{equation} | (3.12) |
which together with Poincaré inequality yields
\begin{equation} \begin{aligned} \| \theta\|_{L^2}\le & C(1+\| \nabla \theta\|_{L^2}). \end{aligned} \end{equation} | (3.13) |
And consequently, (3.11) holds.
Secondly, according to [17,21,33], Multiplying equations (1.1)_{2} by u \theta and integrating by parts yield
\begin{equation} \begin{aligned} &\mu\int \theta| \nabla u|^2dx+(\mu+\lambda)\int \theta| {{\rm{div }}} u|^2dx\\ & = -\int \rho\dot u\cdot \theta u dx-\mu\int u\cdot \nabla \theta \cdot \nabla udx-(\mu+\lambda)\int {{\rm{div }}} u u\cdot \nabla \theta dx \\ &\quad-\int \nabla P\cdot \theta u dx+\int H\cdot \nabla H\cdot \theta u dx-\frac{1}{2}\int \nabla |H|^2 \cdot \theta udx\\ & \triangleq \sum\limits_{i = 1}^6 I_i. \end{aligned} \end{equation} | (3.14) |
Using the same arguments in [17,33], we have
\begin{equation} \begin{aligned} \sum\limits_{i = 1}^4 I_i \le & \eta\| \rho^{1/2} \dot u\|_{L^2}^2+ \varepsilon \| \nabla \theta\|_{L^2}^2 +C\| \rho^{1/2} \theta\|_{L^2}^2\| {{\rm{div }}} u\|_{L^\infty}\\ &+C\left(\| \rho^{1/2} \theta\|_{L^2}^2+\| \nabla u\|_{L^2}^2\right)\|u\|_{L^\infty}^2. \end{aligned} \end{equation} | (3.15) |
Besides, according to (3.1) and (3.11) yields
\begin{equation} \begin{aligned} \sum\limits_{i = 5}^6 I_i\le & C\| \theta\|_{L^{\tilde b}}\|H\|_{L^b}\| \nabla H\|_{L^2}\|u\|_{L^\infty} \le & \varepsilon \| \nabla \theta\|_{L^2}^2+ C \| \nabla H\|_{L^2}^2\|u\|_{L^\infty}^2+C. \end{aligned} \end{equation} | (3.16) |
Substituting (3.10), (3.15) and (3.16) into (3.9), and choosing \varepsilon suitably small, we have
\begin{equation} \begin{aligned} & c_v \frac{d}{dt} \|\rho^{1/2} \theta\|^2_{L^{2}} + \kappa \| \nabla \theta\|_{L^2}^2\\ &\le 2\eta\| \rho^{1/2} \dot u\|_{L^2}^2+C_1\|\Delta H\|_{L^2}^2\\ &\quad+C\left(\| \rho^{1/2} \theta\|_{L^2}^2+\| \nabla u\|_{L^2}^2+\| \nabla H\|_{L^2}^2+1\right)\left(\| {{\rm{div }}} u\|_{L^\infty} + \|u\|_{L^\infty}^2+1\right), \end{aligned} \end{equation} | (3.17) |
where one has applied the key fact as follows
\begin{equation} \begin{aligned} \| \nabla^2H\|_{L^2}\le C \|\Delta H\|_{L^2}. \end{aligned} \end{equation} | (3.18) |
Furthermore, it follows from (3.1) and (1.1) _4 that
\begin{equation} \begin{aligned} & \|H_t\|_{L^2}^2+ \nu^{2}\|\Delta H\|_{L^2}^2+\nu\frac{d}{dt}\| \mbox{curl} H\|_{L^2}^2 \\ &\le C\| \nabla u\|_{L^2}\| \nabla u\|_{L^{2\tilde b}}\|H\|_{L^b}\|H\|_{L^{2\tilde b}} +C\| \nabla H\|_{L^2}^2 \|u\|_{L^\infty}^2\\ &\le C\| \nabla u\|_{L^2} \| \nabla u\|_{L^{2\tilde b}} (\| \nabla H\|_{L^2}+1)+C\| \nabla H\|_{L^2}^2\|u\|_{L^\infty}^2. \end{aligned} \end{equation} | (3.19) |
In order to estimate \| \nabla u\|_{L^{2\tilde b}} , according to [24,26], we divide u into v and w . More precisely, let
\begin{equation} \begin{aligned} u = v+w, \; \; \mbox{and}\; \; v = \mathcal{L}^{-1} \nabla P, \end{aligned} \end{equation} | (3.20) |
then we get
\begin{equation} \begin{aligned} \mathcal{L}w = \rho\dot u-H\cdot \nabla H+\frac{1}{2} \nabla |H|^2. \end{aligned} \end{equation} | (3.21) |
And hence, Lemma 2.2 implies that for any r > 1 ,
\begin{equation} \begin{aligned} \| \nabla v\|_{L^r}\le C\| \theta \rho\|_{L^r}, \end{aligned} \end{equation} | (3.22) |
and
\begin{equation} \begin{aligned} \| \nabla^2 w\|_{L^r}\le C\| \rho\dot u\|_{L^r}+C\||H|| \nabla H|\|_{L^r}. \end{aligned} \end{equation} | (3.23) |
Consequently, it follows from Gagliardo-Nirenberg inequality, (3.2), (3.11), (3.20), (3.22) and (3.23) that for any s\ge 2 ,
\begin{equation} \begin{aligned} \| \nabla u\|_{L^s}\le &C\| \nabla v\|_{L^s}+C\| \nabla w\|_{L^s} \\ \le & C\| \rho \theta\|_{L^s}+C\| \nabla w\|_{L^2}+C\| \nabla w\|_{L^2}^{2/s}\| \nabla^2 w\|_{L^2}^{1-2/s}\\ \le & C\| \rho \theta\|_{L^s}+C\| \nabla w\|_{L^2}+C\| \nabla w\|_{L^2}^{2/s}\left(\| \rho\dot u\|_{L^2}+ \||H|| \nabla H|\|_{L^2}\right)^{1-2/s}\\ \le & \eta\| \rho^{1/2}\dot u\|_{L^2}+C\| \rho \theta\|_{L^s}+C\| \nabla w\|_{L^2} + C\||H|| \nabla H|\|_{L^2} \\ \le &\eta\| \rho^{1/2}\dot u\|_{L^2}+C\| \nabla u\|_{L^2} +C\| \nabla \theta\|_{L^2}+C\| \nabla H\|_{L^2}+ C\|\Delta H\|_{L^2}+C. \end{aligned} \end{equation} | (3.24) |
Putting (3.24) into (3.19) and utilizing Young inequality lead to
\begin{equation} \begin{aligned} &\|H_t\|_{L^2}^2+ \frac{\nu^{2}}{2}\|\Delta H\|_{L^2}^2+\nu\frac{d}{dt}\| \mbox{curl} H\|_{L^2}^2 \\ &\le \varepsilon\| \nabla \theta\|_{L^2}^2+\eta\| \rho^{1/2}\dot u\|_{L^2}^2 +C \left(\| \nabla u\|_{L^2}^2+\|u\|_{L^\infty}^2+1\right) \left(\| \nabla H\|_{L^2}^2+1 \right). \end{aligned} \end{equation} | (3.25) |
Adding (3.25) multiplied by 2\nu^{-2}(C_1+1) to (3.17) and choosing \varepsilon suitably small, we have
\begin{equation} \begin{aligned} &\frac{\kappa}{2} \| \nabla \theta\|_{L^2}^2+2\nu^{-2}(C_1+1)\|H_t\|_{L^2}^2+ \|\Delta H\|_{L^2}^2 \\ &\quad+\frac{d}{dt} \left(c_v \|\rho^{1/2} \theta\|^2_{L^{2}}+2\nu^{-1}(C_1+1)\| \mbox{curl} H\|_{L^2}^2 \right) \\ &\le C\left(\| \rho^{1/2} \theta\|_{L^2}^2+\| \nabla u\|_{L^2}^2+\| \nabla H\|_{L^2}^2+1\right)\left(\| \nabla u\|_{L^2}^2+ \|u\|_{L^\infty}^2+\| {{\rm{div }}} u\|_{L^\infty} +1\right)\\ &\quad+C\eta\| \rho^{1/2}\dot u\|_{L^2}^2. \end{aligned} \end{equation} | (3.26) |
Then, multiplying (1.1) _2 by u_t and integrating by parts, we get
\begin{equation} \begin{aligned} & \frac{1}{2}\frac{d}{dt} \left(\mu\| \nabla u\|^2_{L^{2}}+(\mu+\lambda)\| {{\rm{div }}} u\|_{L^2}^2 \right) + \| \rho^{1/2}\dot u\|_{L^2}^2\\ &\le \eta\| \rho^{1/2}\dot u\|_{L^2}^2 +C\| \nabla u\|_{L^2}^2\|u\|_{L^\infty}^2\\ &\quad+\frac{d}{dt}\left(\int P {{\rm{div }}} u dx+ \frac{1}{2}\int |H|^2 {{\rm{div }}} u dx-\int H\cdot \nabla u \cdot Hdx\right)\\ &\quad-\int P_t {{\rm{div }}} u dx-\int H\cdot H_t {{\rm{div }}} udx +\int H_t\cdot \nabla u\cdot H dx+\int H\cdot \nabla u\cdot H_tdx. \end{aligned} \end{equation} | (3.27) |
Notice that
\begin{equation} \begin{aligned} &\int P_t {{\rm{div }}} u dx = \int P_t {{\rm{div }}} v dx+\int P_t {{\rm{div }}} w dx, \end{aligned} \end{equation} | (3.28) |
integration by parts together with (3.20) leads to
\begin{equation} \begin{aligned} \int P_t {{\rm{div }}} v dx & = \frac{1}{2}\frac{d}{dt}\left((\mu+\lambda)\| {{\rm{div }}} v\|_{L^2}^2+\mu\| \nabla v\|_{L^2}^2\right). \end{aligned} \end{equation} | (3.29) |
Moreover, define
E\triangleq c_v \theta+\frac{1}{2}|u|^2, |
according to (1.1) that E satisfies
\begin{equation} \begin{aligned} (\rho E)_t + {{\rm{div }}}(\rho uE+Pu) = &\Delta \left(\kappa \theta+\frac{1}{2}\mu|u|^2 \right) + \mu {{\rm{div }}}(u\cdot \nabla u)+ \lambda {{\rm{div }}}(u {{\rm{div }}} u)\\ &+H\cdot \nabla H\cdot u-\frac{1}{2}u\cdot \nabla |H|^2+\nu| \mbox{curl} H|^2. \end{aligned} \end{equation} | (3.30) |
Motivated by [17,21], it can be deduced from (3.30) that
\begin{equation} \begin{aligned} &-\int P_t {{\rm{div }}} w dx\\ & = -\frac{R}{c_v} \left(\int ( \rho E)_t {{\rm{div }}} wdx-\int \frac{1}{2}( \rho |u|^2)_t {{\rm{div }}} wdx \right)\\ & = -\frac{R}{c_v} \left\{\int \left( (c_v+R) \rho \theta u+\frac{1}{2} \rho|u|^2 u-\kappa \nabla \theta-\mu \nabla u\cdot u-\mu u\cdot \nabla u-\lambda u {{\rm{div }}} u \right)\cdot \nabla {{\rm{div }}} wdx \right.\\ &\quad\quad -\frac{1}{2}\int \rho|u|^2 u \cdot \nabla {{\rm{div }}} wdx-\int \rho \dot u\cdot u {{\rm{div }}} wdx\\ &\quad\quad -\int {{\rm{div }}} H H\cdot u {{\rm{div }}} wdx-\int H\cdot \nabla u\cdot H {{\rm{div }}} wdx-\int (H\cdot u) H\cdot \nabla {{\rm{div }}} wdx\\ &\quad\quad +\frac{1}{2}\int {{\rm{div }}} u |H|^2 {{\rm{div }}} wdx+\frac{1}{2}\int |H|^2 u \cdot \nabla {{\rm{div }}} wdx\\ &\quad\quad \left.-\nu\int \nabla {{\rm{div }}} w \times \mbox{curl} H \cdot H dx-\nu\int \mbox{curl}( \mbox{curl} H )\cdot H {{\rm{div }}} wdx \right\}\\ &\le C \eta\| \rho^{1/2}\dot u\|_{L^2}^2+C\| \nabla \theta\|_{L^2}^2 +C\|\Delta H\|_{L^2}^2\\ &\quad+C \left(\| \nabla u\|_{L^2}^2+\|u\|_{L^\infty}^2+1 \right) \left(\| \rho^{1/2} \theta\|_{L^2}^2+\| \nabla u\|_{L^2}^2+\| \nabla H\|_{L^2}^2+1 \right). \end{aligned} \end{equation} | (3.31) |
Additionally, combining (3.1) and (3.24) yields
\begin{equation} \begin{aligned} &\int H_t\cdot \nabla u\cdot H dx+\int H\cdot \nabla u\cdot H_tdx-\int H\cdot H_t {{\rm{div }}} udx \\ &\le C\|H_t\|_{L^2}^2+C\| \nabla u\|_{L^{\tilde b}}^2\|H\|_{L^b}^2\\ &\le C\eta\| \rho^{1/2}\dot u\|_{L^2}^2 +C\left( \|H_t\|_{L^2}^2+ \| \nabla \theta\|_{L^2}^2+ \| \nabla u\|_{L^2}^2 + \| \nabla H\|_{L^2}^2+ \|\Delta H\|_{L^2}^2+1\right). \end{aligned} \end{equation} | (3.32) |
Substituting (3.28), (3.29), (3.31) and (3.32) into (3.27) yields
\begin{equation} \begin{aligned} & \| \rho^{1/2}\dot u\|_{L^2}^2+ \frac{d}{dt} \left(\frac{\mu}{2}\left(\| \nabla u\|^2_{L^{2}}+\| \nabla v\|^2_{L^{2}}\right)+\frac{ \mu+\lambda }{2}\left(\| {{\rm{div }}} u\|_{L^2}^2 +\| {{\rm{div }}} v\|_{L^2}^2\right)-A(t) \right) \\ &\le C_2\left(\| \nabla \theta\|_{L^2}^2 +\|H_t\|_{L^2}^2+\|\Delta H\|_{L^2}^2\right)+C\eta\| \rho^{1/2}\dot u\|_{L^2}^2 \\ &\quad+C \left(\| \nabla u\|_{L^2}^2+\|u\|_{L^\infty}^2+1 \right) \left(\| \nabla u\|_{L^2}^2+\| \nabla H\|_{L^2}^2+\| \rho^{1/2} \theta\|_{L^2}^2+1 \right) , \end{aligned} \end{equation} | (3.33) |
where
\begin{equation} \begin{aligned} & A(t)\triangleq \frac{1}{2}\int |H|^2 {{\rm{div }}} u dx+\int P {{\rm{div }}} u dx-\int H\cdot \nabla u \cdot Hdx, \end{aligned} \end{equation} | (3.34) |
satisfies
\begin{equation} \begin{aligned} A(t)\le \frac{\mu}{4}\| \nabla u\|_{L^2}^2+C_3 \left(\| \rho^{1/2} \theta\|_{L^2}^2+ \| \mbox{curl} H\|_{L^2}^2+1 \right). \end{aligned} \end{equation} | (3.35) |
Recalling the inequality (3.26), let
\begin{equation} \begin{aligned} C_4 = \min\{2\nu^{-2}(C_1+1), \; \frac{\kappa}{2}, \; 1\}, \; \; \; C_5 = \min\{2\nu^{-1}(C_1+1), \; c_v\}, \end{aligned} \end{equation} | (3.36) |
adding (3.26) multiplied by C_6 = \max\{C_4^{-1}(C_2+1), C_5^{-1}(C_3+1)\} into (3.33) and choosing \eta suitably small, we have
\begin{equation} \begin{aligned} &\frac{d}{dt}\tilde{A}(t)+ \frac{1}{2}\| \rho^{1/2}\dot u\|_{L^2}^2+\| \nabla \theta\|_{L^2}^2+ \|H_t\|_{L^2}^2+ \|\Delta H\|_{L^2}^2 \\ &\le C\left(\| \rho^{1/2} \theta\|_{L^2}^2+\| \nabla u\|_{L^2}^2+\| \nabla H\|_{L^2}^2+1\right)\left(\| \nabla u\|_{L^2}^2 + \|u\|_{L^\infty}^2+\| {{\rm{div }}} u\|_{L^\infty} +1\right), \end{aligned} \end{equation} | (3.37) |
where
\begin{equation} \begin{aligned} \tilde{A}(t)\triangleq&C_6\left( c_v \|\rho^{1/2} \theta\|^2_{L^{2}}+2\nu^{-1}(C_1+1)\| \mbox{curl} H\|_{L^2}^2\right)\\ &+\frac{\mu}{2}(\| \nabla u\|^2_{L^{2}}+\| \nabla v\|^2_{L^{2}})+\frac{ \mu+\lambda }{2}(\| {{\rm{div }}} u\|_{L^2}^2+\| {{\rm{div }}} v\|_{L^2}^2) -A(t), \end{aligned} \end{equation} | (3.38) |
satisfies
\begin{equation} \begin{aligned} &\|\rho^{1/2} \theta\|^2_{L^{2}}+\frac{\mu}{4}\| \nabla u\|^2_{L^{2}} +\| \mbox{curl} H\|_{L^2}^2-C \\ &\le \tilde{A}(t) \le C \| \rho^{1/2} \theta\|_{L^2}^2+C\| \nabla u\|_{L^2}^2+C \| \mbox{curl} H\|_{L^2}^2+C. \end{aligned} \end{equation} | (3.39) |
Finally, integrating (3.37) over (\tau, t) , along with (3.39) yields
\begin{equation} \begin{aligned} \psi(t)\le C\int_\tau^t\left(\| \nabla u\|_{L^2}^2 + \|u\|_{L^\infty}^2+\| {{\rm{div }}} u\|_{L^\infty} +1\right)\psi(s)ds+C \psi(\tau), \end{aligned} \end{equation} | (3.40) |
where
\begin{equation} \begin{aligned} \psi(t)\triangleq &\int_0^t\left(\| \rho^{1/2}\dot u\|_{L^2}^2+\| \nabla \theta\|_{L^2}^2+ \|H_t\|_{L^2}^2+ \|\Delta H\|_{L^2}^2 \right)ds\\ &+\|\rho^{1/2} \theta\|^2_{L^{2}}+ \| \nabla u\|^2_{L^{2}}+ \| \mbox{curl} H\|_{L^2}^2+1. \end{aligned} \end{equation} | (3.41) |
Combined with (3.1), (3.3) and Gronwall inequality implies that for any 0 < \tau\le t < T^* ,
\begin{equation} \begin{aligned} \psi(t)&\le C \psi(\tau)\exp\left\{ \int_\tau^t\left(\| \nabla u\|_{L^2}^2 + \|u\|_{L^\infty}^2+\| {{\rm{div }}} u\|_{L^\infty} +1\right)ds\right\}\\ &\le C \psi(\tau) \exp\left\{\int_\tau^t\|u \|_{L^\infty}^2 ds\right\}. \end{aligned} \end{equation} | (3.42) |
Utilizing Lemma 2.5, we have
\begin{equation} \begin{aligned} \|u\|_{L^2(\tau, t;L^\infty)}^2 \le C \| u \|_{L^2(\tau, t;H^1)}^2 \ln\left(e+ \|u\|_{L^2(\tau, t; W^{1, b})} \right)+C. \end{aligned} \end{equation} | (3.43) |
Combining (3.1), (3.2), (3.11), (3.22), (3.23) and Sobolev inequality leads to
\begin{equation} \begin{aligned} \|u\|_{W^{1, b}} &\le \|v\|_{W^{1, b}}+C\|w\|_{W^{2, 2b/(b+2)}}\\ &\le C\| \rho\dot u\|_{L^{2b/(b+2)}}+C\| \rho \theta\|_{L^b}+C\|u\|_{L^2}+C\||H|| \nabla H|\|_{L^{2b/(b+2)}}\\ &\le C\| \rho^{1/2}\|_{L^b}\| \rho^{1/2}\dot u\|_{L^2}+C\| \nabla \theta\|_{L^2}+C\| \nabla u\|_{L^2}+C\|H\|_{L^b}\| \nabla H\|_{L^2}+C\\ &\le C\| \rho^{1/2}\dot u\|_{L^2}+C\| \nabla \theta\|_{L^2}+C\| \nabla u\|_{L^2}+C\| \nabla H\|_{L^2}+C, \end{aligned} \end{equation} | (3.44) |
this implies that
\begin{equation} \begin{aligned} \|u\|_{L^2(\tau, t; W^{1, b})} \le &C\psi^{1/2}(t). \end{aligned} \end{equation} | (3.45) |
Substituting (3.45) into (3.43) indicates
\begin{equation} \begin{aligned} \|u\|_{L^2(\tau, t;L^\infty)}^2 \le C+C \| u \|_{L^2(\tau, t;H^1)}^2 \ln\left(C\psi (t)\right)\le C+\ln\left(C\psi(t)\right)^{C_7\| u \|_{L^2(\tau, t;H^1)}^2}. \end{aligned} \end{equation} | (3.46) |
Using (3.3), one can choose some \tau which is close enough to t such that
\begin{equation} \begin{aligned} C_7\| u \|_{L^2(\tau, t;H^1)}^2\le \frac{1}{2}, \end{aligned} \end{equation} | (3.47) |
which together with (3.42) and (3.46) yields
\begin{equation} \begin{aligned} \psi(t)& \le C \psi^{2}(\tau) \le C. \end{aligned} \end{equation} | (3.48) |
Noticing the definition of \psi in (3.41), we immediately have (3.8). The proof of Lemma 3.3 is completed.
Now, we show some higher order estimates of the solutions which are needed to guarantee the extension of local solution to be a global one under the conditions (1.12)–(1.14) and (3.1).
Lemma 3.4. Under the assumptions of Theorem 1.2 and (3.1), it holds that for any t\in [0, T^*),
\begin{equation} \begin{aligned} &\sup\limits_{0\le s\le t} \left(\| \rho\|_{W^{1, q}} + \| \theta\|_{H^2}+\|u\|_{H^2}+\|H\|_{H^2}\right) \le C. \end{aligned} \end{equation} | (3.49) |
Proof. First, it follows from (3.8), Gagliardo-Nirenberg and Poincaré inequalities that for 2\le q < \infty ,
\begin{equation} \begin{aligned} \|u\|_{L^q}+\|H\|_{L^q}\le C. \end{aligned} \end{equation} | (3.50) |
Combining (1.1) _4 , (3.3), (3.8) and (3.18) yields
\begin{equation} \begin{aligned} \|H\|_{H^2} +\| \nabla H\|_{L^4}^2 &\le C \| \nabla u\|_{L^4}+C\|H_t\|_{L^2} +C. \end{aligned} \end{equation} | (3.51) |
Furthermore, it can be deduced from (3.8), (3.24), (3.50) and (3.51) that
\begin{equation} \begin{aligned} \| \nabla u\|_{L^4} &\le C \| \rho^{1/2} \dot u\|_{L^2}+C \| \nabla \theta\|_{L^2} +C\|H_t\|_{L^2} +C . \end{aligned} \end{equation} | (3.52) |
Then, according to (3.11) and Sobolev inequality, we get
\begin{equation} \begin{aligned} \| \theta\|_{L^\infty}^2\le \varepsilon\| \nabla^2 \theta\|_{L^2}^2+C\| \nabla \theta\|_{L^2}^2+C, \end{aligned} \end{equation} | (3.53) |
which combined with (1.1) _3 , (3.8), and choosing \varepsilon suitably small yield
\begin{equation} \begin{aligned} \| \theta\|_{H^2}^2 \le C\| \rho^{1/2}\dot \theta\|_{L^2}^2 +C\| \nabla \theta\|_{L^2}^2+C\| \nabla u\|_{L^4}^4+C\| \nabla H\|_{L^4}^4+C. \end{aligned} \end{equation} | (3.54) |
Therefore, the combination of (3.51) and (3.52) yields
\begin{equation} \begin{aligned} &\sup\limits_{0\le s\le t} \left( \| \theta\|_{L^r} + \| \nabla \theta\|_{L^2}+ \| \nabla u\|_{L^4} + \|H\|_{H^2} +\| \nabla H\|_{L^4} \right) \le C, \; \; \; \; \forall\; r\ge 1. \end{aligned} \end{equation} | (3.55) |
Together with (3.53) and (3.54) gives
\begin{equation} \begin{aligned} \sup\limits_{0\le s\le t} \left(\| \theta\|_{H^2} +\| \theta\|_{L^\infty} \right)\le C. \end{aligned} \end{equation} | (3.56) |
Now, we bound \| \nabla \rho\|_{W^{1, q}} and \|u\|_{H^2} . For r\in [2, q], it holds that
\begin{equation} \begin{aligned} \frac{d}{dt}\|\nabla\rho\|_{L^r}&\leq C\|\nabla\rho\|_{L^r}(\|\nabla u \|_{L^\infty}+1)+C\| \nabla^2u\|_{L^r} \\ &\le C\|\nabla\rho\|_{L^r}(\|\nabla v \|_{L^\infty}+\|\nabla w \|_{L^\infty}+1)+C\| \nabla^2v\|_{L^r} +C\| \nabla^2w\|_{L^r} \\ &\le C\|\nabla\rho\|_{L^r}(\|\nabla v \|_{L^\infty}+\|\nabla w \|_{L^\infty}+1)+C\| \nabla^2w\|_{L^r} +C, \end{aligned} \end{equation} | (3.57) |
where in the last inequality one has applied the following fact
\begin{equation} \begin{aligned} \| \nabla^2v\|_{L^r} \le C\| \nabla \rho \|_{L^r}+C. \end{aligned} \end{equation} | (3.58) |
Taking (3.2), (3.56), (3.58) and Lemmas 2.2–2.4, we get
\begin{equation} \begin{aligned} \| \nabla v\|_{L^\infty}\le C \ln(e+\| \nabla \rho\|_{L^r})+C. \end{aligned} \end{equation} | (3.59) |
Putting (3.59) into (3.57), it can be deduced from Gronwall inequality that
\begin{equation} \begin{aligned} \| \nabla \rho\|_{L^r}&\leq C. \end{aligned} \end{equation} | (3.60) |
Finally, let r = 2 in (3.60), according to Lemma 2.2, (3.50), (3.55) and (3.58) yields
\begin{equation} \begin{aligned} \|u\|_{H^2} &\le C. \end{aligned} \end{equation} | (3.61) |
Therefore, together with (3.55), (3.56), (3.60) and (3.61), we get (3.49). The proof of Lemma 3.4 is completed.
With the priori estimates in Lemmas 3.1–3.4, we can prove Theorem 1.2.
Proof of Theorem 1.2. Assume that (1.16) is false, namely, (3.1) holds. Notice that the general constant C in Lemmas 3.1–3.4 is independent of t , that is, all the priori estimates attained in Lemmas 3.1–3.4 are uniformly bounded for any t\le T^* . Therefore, the function
( \rho, \theta, u, H)(x, T^*)\triangleq \lim\limits_{t\rightarrow T^*}( \rho, \theta, u, H)(x, t) |
satisfies the initial conditions (1.12) at t = T^* .
Due to
( \rho\dot u, \rho\dot \theta)(x, T^*) = \lim\limits_{t\rightarrow T^*}( \rho\dot u, \rho\dot \theta)\in L^2, |
therefore
\begin{equation} \notag \begin{aligned} - \mu{\triangle} u - (\mu + {\lambda})\nabla {{\rm{div }}} u + R\nabla ( \rho \theta )-H \cdot \nabla H +\frac{1}{2} \nabla |H |^2|_{t = T^*} & = \rho^{1/2}(x, T^*)g_1(x), \\ - \kappa{\triangle} \theta - 2\mu|\mathfrak{D}(u )|^2-\lambda( {{\rm{div }}} u )^2-\nu( \mbox{curl} H)^2|_{t = T^*}& = \rho^{1/2}(x, T^*)g_2(x), \end{aligned} \end{equation} |
with
\begin{equation} \notag \begin{aligned} g_1(x)\triangleq\begin{cases} \rho^{-1/2}(x, T^*)( \rho\dot u)(x, T^*), \; \; \; \; &\mbox{for}\; x\in\{x| \rho(x, T^*) > 0\}, \\0, \; \; \; \; &\mbox{for}\; x\in\{x| \rho(x, T^*) = 0\}, \end{cases} \end{aligned} \end{equation} |
and
\begin{equation} \notag \begin{aligned} g_2(x)\triangleq\begin{cases} \rho^{-1/2}(x, T^*)(c_v \rho\dot \theta+R \theta \rho {{\rm{div }}} u)(x, T^*), \; \; \; \; &\mbox{for}\; x\in\{x| \rho(x, T^*) > 0\}, \\0, \; \; \; \; &\mbox{for}\; x\in\{x| \rho(x, T^*) = 0\}, \end{cases} \end{aligned} \end{equation} |
satisfying g_1, \; g_2\in L^2 . Thus, (\rho, \theta, u, H)(x, T^*) also satisfies (1.13) and (1.14).
Hence, Theorem 1.1 shows that we could extend the local strong solutions beyond T^* , while taking (\rho, \theta, u, H)(x, T^*) as the initial data. This contradicts the hypothesis of Theorem 1.2 that T^* is the maximum existence time of the strong solution. This completes the proof of theorem 1.2.
This paper concerns the blow-up criterion for the initial boundary value problem of the two-dimensional full compressible magnetohydrodynamic equations in the Eulerian coordinates. When the initial density allowed to vanish, and the magnetic field H satisfies the perfect conducting boundary condition H\cdot n = \mbox{curl} H = 0 , we prove the blow-up criterion \lim\limits_{t\rightarrow T^*}\big(\|H\|_{L^\infty(0, t; L^b)}+\| {{\rm{div }}} u\|_{L^1(0, t; L^\infty)}\big) = \infty for any b > 2 , which depending on both H and {{\rm{div }}} u .
The author sincerely thanks the editors and anonymous reviewers for their insightful comments and constructive suggestions, which greatly improved the quality of the paper. The research was partially supported by the National Natural Science Foundation of China (No.11971217).
The author declares no conflict of interest in this paper.
[1] |
S. Hussain, H. Arif, M. Noorullah, A. Athanasios, Pantelous, pricing American options under Azzalini Ito-McKean skew Brownian motions, Appl. Math. Comput., 451 (2023), 128040. https://doi.org/10.1016/j.amc.2023.128040 doi: 10.1016/j.amc.2023.128040
![]() |
[2] |
M. Shirzadi, M. Rostami, M. Dehghan, X. Li, American options pricing under regime-switching jump-diffusion models with meshfree finite point method, Chaos Solitons Fract., 166 (2023), 112919. https://doi.org/10.1016/j.chaos.2022.112919 doi: 10.1016/j.chaos.2022.112919
![]() |
[3] |
H. Song, J. Xu, J. Yang, Y. Li, Projection and contraction method for the valuation of American options under regime switching, Commun. Nonlinear Sci. Numer. Simul., 109 (2022), 106332. https://doi.org/10.1016/j.cnsns.2022.106332 doi: 10.1016/j.cnsns.2022.106332
![]() |
[4] |
J. Li, C. Bi, Study of weak solutions of variational inequality systems with degenerate parabolic operators and quasilinear terms arising Americian option pricing problems, AIMS Math., 7 (2022), 19758–19769. https://doi.org/10.3934/math.20221083 doi: 10.3934/math.20221083
![]() |
[5] |
T. Wu, Some results for a variation-inequality problem with fourth order p(x)-Kirchhoff operator arising from options on fresh agricultural products, AIMS Math., 8 (2023), 6749–6762. https://doi.org/10.3934/math.2023343 doi: 10.3934/math.2023343
![]() |
[6] |
C. O. Alves, L. M. Barros, C. E. T. Ledesma, Existence of solution for a class of variational inequality in whole RN with critical growth, J. Math. Anal. Appl., 494 (2021), 124672. https://doi.org/10.1016/j.jmaa.2020.124672 doi: 10.1016/j.jmaa.2020.124672
![]() |
[7] |
Y. Bai, S. Migorski, S. Zeng, A class of generalized mixed variational-hemivariational inequalities Ⅰ: Existence and uniqueness results, Comput. Math. Appl., 79 (2020), 2897–2911. https://doi.org/10.1016/j.camwa.2019.12.025 doi: 10.1016/j.camwa.2019.12.025
![]() |
[8] |
Y. Wang, C. Zhang, Existence results of partial differential mixed variational inequalities without Lipschitz continuity, J. Math. Anal. Appl., 484 (2020), 123710. https://doi.org/10.1016/j.jmaa.2019.123710 doi: 10.1016/j.jmaa.2019.123710
![]() |
[9] |
M. T. O. Pimenta, R. Servadei, Some existence results for variational inequalities with nonlocal fractional operators, Nonlinear Anal., 189 (2019), 111561. https://doi.org/10.1016/j.na.2019.06.020 doi: 10.1016/j.na.2019.06.020
![]() |
[10] |
W. Han, Y. Li, Stability analysis of stationary variational and hemivariational inequalities with applications, Nonlinear Anal. Real World Appl., 50 (2019), 171–191. https://doi.org/10.1016/j.nonrwa.2019.04.009 doi: 10.1016/j.nonrwa.2019.04.009
![]() |
[11] |
J. Fan, R. Zhong, Stability analysis for variational inequality in reflexive Banach spaces, Nonlinear Anal. Theor., 69 (2008), 2566–2574. https://doi.org/10.1016/j.na.2007.08.031 doi: 10.1016/j.na.2007.08.031
![]() |
[12] |
Y. Feng, Regularity of weak solutions to a class of fourth order parabolic variational inequality problems arising from swap option pricing, AIMS Math., 8 (2023), 13889–13897. https://doi.org/10.3934/math.2023710 doi: 10.3934/math.2023710
![]() |
[13] |
M. Sofiani, On parabolic partial differential equations with Hölder continuous diffusion coefficients, J. Math. Anal. Appl., 526 (2023), 127224. https://doi.org/10.1016/j.jmaa.2023.127224 doi: 10.1016/j.jmaa.2023.127224
![]() |
[14] | Z. Wu, J. Yin, C. Wang, Elliptic & Parabolic Equations, Singapore: World Scientific Publishing, 2006. |