数学物理学报  2018, Vol. 38 Issue (2): 284-290   PDF    
扩展功能
加入收藏夹
复制引文信息
加入引用管理器
Email Alert
RSS
本文作者相关文章
邢超
任猛章
罗宏
热盐环流方程全局弱解的存在性
邢超1, 任猛章2, 罗宏1     
1. 四川师范大学数学与软件科学学院 成都 610066;
2. 四川民族学院数学系 四川康定 626001
摘要:该文利用T-弱连续算子理论和空间序列方法证明了热盐环流方程全局弱解的存在性.首先根据热盐环流方程的形式选择试探函数空间和解函数空间,再将方程化为抽象的算子方程,验证算子是T-弱连续的并满足对应条件,从而得到热盐环流方程全局弱解的存在性.
关键词全局弱解    热盐环流方程    存在性    
The Existence of Global Weak Solutions to Thermohaline Circulation Equations
Xing Chao1, Ren Mengzhang2, Luo Hong1     
1. College of Mathematics and Software Science, Sichuan Normal University, Chengdu 610066;
2. Mathematic Department, Sichuan Minzu College, Sichuan Kangding 626001
Abstract: In the article, the existence of global weak solutions to thermohaline circulation equations is proved by the theory of T-weakly continuous operator and the method of space sequence. Firstly, the space of test function and space of solution function are chosen according to thermohaline circulation equations. Then the equations can be rewritten as abstract operator equation. Furthermore, it is proved that the operator is T-weakly continuous and satisfies corresponding conditions. Thus the existence of global weak solutions to thermohaline circulation equations is obtained.
Key words: Global weak solutions     Thermohaline circulation equations     Existence    
1 引言

热盐环流是气候改变的重要内在原因, 同时也是地球上最大的海洋环流.热盐海洋环流已持续数十年之久.在物理观测和数值分析等方面已有大量的研究[1-7].

描述热盐环流运动状态的数学模型是如下的方程[7]

$ \begin{equation} \label{eqa28} \frac{\partial u}{\partial t}=P_{r}(\Delta u -\nabla p)+P_{r}(RT-\textrm{sign}(S_{0}-S_{1})\tilde{R}S)\vec{k}-(u\cdot\nabla)u, \end{equation} $ (1.1)
$ \begin{equation} \label{eqa29} \frac{\partial T}{\partial t}=\Delta T+u_{3}-(u\cdot \nabla)T+Q, \end{equation} $ (1.2)
$ \begin{equation} \label{eqa30} \frac{\partial S}{\partial t}=L_{e}\Delta S+\textrm{sign}(S_{0}-S_{1})u_{3}-(u\cdot \nabla)S, \end{equation} $ (1.3)
$ \begin{equation} \label{eqa31}\textrm{div} u=0, \end{equation} $ (1.4)

其中$(x, t)=(x_1, x_2, x_{3}, t) \in \Omega \times (0, \infty)$; 未知函数$u=(u_{1}, u_{2}, u_{3})$, $T$, $S$, $p$, 分别表示速度场, 温度, 盐度, 以及压力; $Q(x)$是热源, 是已知函数; $\vec{k}=(0, 0, 1)$; 参数$P_{r}$, $R$, $\tilde{R}$, $L_{e}$, $S_{0}$, $S_{1}$都是正常数; 以及区域$\Omega=D\times(0, 1)$是$\mathbb{R}^{3}$的有界开集, $D=\{(x_{1}, x_{2})|x_{1}^{2}+x_{2}^{2} < r^{2}\}$是圆形区域.

对于问题(1.1)-(1.4), 考虑如下的边值问题

$ \begin{equation} \label{eqa05} u_{n}=0, \frac{\partial u_{3}}{\partial n}=\frac{\partial T}{\partial n}=\frac{\partial S}{\partial n}=0, \ \ \ \ x\in \partial D\times(0, 1), \end{equation} $ (1.5)
$ \begin{equation} \label{eqa06} u_{3}=T=S=0, \frac{\partial u_{1}}{\partial x_{3}}=\frac{\partial u_{2}}{\partial x_{3}}=0, \ \ \ x_{3}=0, 1, \end{equation} $ (1.6)

其中$u_{n}$, $u_{3}$分别表示速度场在柱体区域侧边界的法向分量和切向分量.同时考虑如下的初始条件

$ \begin{eqnarray} \label{eqa07} (u(0), T(0), S(0))=(\tilde{u_{0}}, \tilde{T_{0}}, \tilde{S_{0}}). \end{eqnarray} $ (1.7)

热盐环流方程(1.1)-(1.4)是Navier-Stokes方程耦合温度和盐度的方程.马天和汪守宏[7]通过无量纲化得到方程(1.1)-(1.4)并得到了热盐环流发生跃迁的条件.因为特征值问题, 他们考虑的是方体区域.因为忽略了热源, 他们给出了方程平凡的稳态解.然而热源是热盐环流的重要因素, 我们[8]利用弱连续算子的锐角原理研究了带热源的热盐环流方程稳态解的存在性, 利用ADN定理研究了解的正则性.故该文考虑了热源.此外, 因热盐环流是周期性运动, 故本文考虑的是管道区域$\Omega=D\times(0, 1)$, 其中$D$是海洋的一部分, [0, 1]是某周期的一部分.

演化方程全局弱解的存在性是研究全局吸引子存在性的前提, 可为研究数值解提供有价值的参考.因此, 存在大量文献研究全局弱解的存在性[9-14].黄海洋和郭柏灵[9]给出了适用于天气预报的大气数学模型, 利用Galerkin方法研究了方程弱解的存在性, 并研究了该模型轨道吸引子的存在性.该文应用T -弱连续算子方法研究了热盐环流方程全局弱解的存在性. T -弱连续算子方法是建立在Galerkin方法和Arzela-Ascoli定理之上, 是将方程的试探函数空间和解函数空间分为两个不同的空间.根据方程, 建立试探函数空间和解函数空间, 利用弱解的定义建立相应地算子.进而算子只需满足某些条件, 便可得到方程解的存在性.该方法对比Galerkin方法而言, 计算过程简单, 同时该方法还适应许多类演化方程全局弱解的存在性的判定.

该文研究了热盐环流方程(1.1)-(1.7)全局弱解的存在性.结构如下:第2部分陈述预备知识; 第3部分应用T-弱连续算子理论证明了全局弱解的存在性.

2 预备知识

该部分将陈述T -弱连续算子理论的相关知识[15].

设$X$是线性空间, $X_{1}, X_{2}$是可分的Banach空间, $H$是Hilbert空间. $X_{1}, X_{2}$, $H$是$X$在各自范数下的完备化. $X_{1}, X_{2}\subset H$是稠密嵌入. $F: X_{2}\times (0, \infty)\rightarrow X_{1}^{*}$是连续映射.考虑如下的抽象方程

$ \begin{eqnarray} \label{eqb21} \left\{ \begin{array}{ll} \frac{{\rm d}u}{{\rm d}t}=Fu, 0 <t<\infty, \\[2mm] u(0)=\varphi, \end{array} \right. \end{eqnarray} $ (2.1)

其中$\varphi\in H$, 以及$u:[0, +\infty)\rightarrow H$是未知函数.

定义2.1[15]   设$\varphi\in H$是给定的初始条件. $u\in L^{p}((0, T), X_{2})\bigcap L^{\infty}((0, T), H), $ $(0 < T < \infty)$称为方程(2.1)的全局弱解, 如果$u$满足

$ \begin{eqnarray} \label{eqb22} (u(t), v)_{H}=\int^{t}_{0}\langle Fu, v\rangle {\rm d}t+(\varphi, v)_{H}, \ \ \forall v\in X_{1}\subset H. \end{eqnarray} $ (2.2)

定义2.2[15]   设$u_{n}, u_{0}\in L^{p}((0, T), X_{2}).$ $u_{n}\rightharpoonup u_{0}$在$L^{p}((0, T), X_{2})$中称为是一致弱收敛的, 如果${u_{n}}\subset L^{\infty}((0, T), H)$是有界的, 而且

$ \begin{eqnarray} \label{eqb23} \left\{ \begin{array}{ll} u_{n}\rightharpoonup u_{0}, \ \ \textrm{在} \ \ L^{p}((0, T), X_{2}), \\[2mm] \lim\limits_{n\rightarrow \infty}\int^{T}_{0}|\langle u_{n}-u_{0}, v\rangle_{H}|^{2}{\rm d}t=0, \forall v\in H. \end{array} \right. \end{eqnarray} $ (2.3)

定义2.3[15]   映射$F: X_{2}\times(0, \infty)\rightarrow X^{*}_{1}$称为是T -弱连续的, 如果对$p=(p_{1}, \cdots, p_{m}), $ $0 < T < \infty$和$u_{n}$在条件(2.3)下一致弱收敛于$u_{0}$, 有

$ \begin{eqnarray} \lim\limits_{n\rightarrow \infty}\int^{T}_{0}\langle Fu_{n}, v\rangle {\rm d}t =\int^{T}_{0}\langle Fu_0, v\rangle {\rm d}t, \ \ \forall v\in X_{1}. \end{eqnarray} $ (2.4)

引理2.1[15]  设$\{u_{n}\}\subset L^{p}((0, T), W^{m, p})(m\geq 1)$是有界序列, 而且$\{u_{n}\}$一致弱收敛于$u_{0}\in L^{p}((0, T), W^{m, p})$, 也就是在$L^{p}((0, T), X_{2})$中的有界序列, 若$\{u_{n}\}\subset L^{\infty}((0, T), H)$是有界的, 而且

$ \begin{eqnarray} \label{eqb4} \left\{ \begin{array}{ll} u_{n}\rightharpoonup u_{0}, \ \mbox{在$ L^{p}((0, T), X_{2})$中, } \\[2mm] \lim\limits_{n\rightarrow \infty}\int^{T}_{0} \bigg[\int_{\Omega}( u_{n}-u_{0})v{\rm d}x \bigg]^{2}{\rm d}t=0, \forall v\in C^{\infty}_{0}. \end{array} \right. \end{eqnarray} $ (2.5)

那么, 对所有的$|\alpha|\leq m-1$, 有

$ \begin{eqnarray} D^{\alpha}u_{n}\rightarrow D^{\alpha}u_{0}, \ \ L^{2}((0, T)\times \Omega). \end{eqnarray} $ (2.6)

引理2.2[15]  假设$F :X_{2}\times (0, \infty)\rightarrow X^{*}_{1}$是T -弱连续, 而且满足如下假设条件

(A1) 存在$p=(p_{1}, \cdots, p_{m}), p_{i}(1\leq i\leq m)$, 使得

$ \begin{eqnarray} \langle Fu, u\rangle \leq -C_{1}\| u\| ^{p}_{X_{2}}+C_{2}\| u\| ^{2}_{H}+f(t), \ \ \forall u\in X, \end{eqnarray} $ (2.7)

其中$C_{1}, C_{2}$是常数, $f\in L^{1}(0, T)(0 < T < \infty), \| \cdot\| ^{p}_{X_{2}}=\sum\limits_{ i=1}^{m}|\cdot|_{i}^{p_{i}}$, $|\cdot|_{i}$是空间$X_{2}$的半范, $\| \cdot\| _{X_{2}}=\sum \limits_{i=1}^{m}|\cdot|_{i}$;

(A2) 存在对所有的$0 < h < 1$, $0 < \alpha < 1$, 以及$u\in C^{1}([0, \infty), X)$, 使得

$ \begin{eqnarray} \bigg|\int^{t+h}_{t}\langle Fu, v\rangle {\rm d}t \bigg|\leq Ch^{\alpha}, \ \ \forall v\in X \hbox{且} \ \ 0\leq t\leq T, \end{eqnarray} $ (2.8)

其中$C>0$只依赖于$T, \| v\| _{X_{1}}, \int^{t}_{0}\| u\| ^{p}_{X_{2}}{\rm d}t$和$\sup\limits_{0\leq t\leq T}\| u\| _{H}$.

则, 对所有的$\varphi\in H$, 方程(2.1)存在全局弱解

$ \begin{eqnarray} u\in L^{\infty}((0, T), H)\bigcap L^{p}((0, T), X_{2}), \ \ 0<T<\infty, \ \ \mbox{$p$ 如条件(A1)}. \end{eqnarray} $ (2.9)
3 全局弱解的存在性

引入如下的空间

$ X=\{\Phi=(u, T, S) \in C^\infty(\Omega, \mathbb{R}^5)| \Phi\ \ \hbox{满足 (1.4)-(1.6)式} \}, $
$ H=\{\Phi=(u, T, S) \in L^2(\Omega, \mathbb{R}^5)| \Phi \ \ \hbox{满足 (1.4)-(1.6)式} \}, $

其中范数为$\| \Phi\| _{H}=[\| u\| ^{2}_{L^{2}(\Omega, \mathbb{R}^3)}+\| T\| ^{2}_{L^{2}(\Omega)}+\| S\| ^{2}_{L^{2}(\Omega)}]^{\frac{1}{2}}$.

$ H_1=\{\Phi=(u, T, S) \in H^1(\Omega, \mathbb{R}^5)| \Phi \ \ \hbox{满足 (1.4)-(1.6)式} \}, $

其中范数为$\| \Phi\| _{H_{1}}=[\| \nabla u\| ^{2}_{L^{2}(\Omega, \mathbb{R}^3)}+\| \nabla T\| ^{2}_{L^{2}(\Omega)}+\| \nabla S\| ^{2}_{L^{2}(\Omega)}]^{\frac{1}{2}}$.

定理3.1  设$\Phi_{0}=(\tilde{u_{0}}, \tilde{T_{0}}, \tilde{S_{0}})\in H, Q\in L^{2}(\Omega)$, 则方程(1.1)-(1.7)存在全局弱解

$ \begin{eqnarray} \label{eqc1} \Phi=(u, T, S)\in L^{\infty}_{\rm Loc}((0, \infty), H)\bigcap L^{2}_{\rm Loc}((0, \infty), H_{1}). \end{eqnarray} $ (3.1)

  对任意的$\Psi=(v, W, Z) \in H_1$, 定义映射$F: H_{1}\times (0, \infty)\rightarrow H_{1}^{*}$,

$ \begin{eqnarray*} \langle F \Phi, \Psi\rangle& =&\int_{\Omega}[-P_{r}\nabla u\cdot\nabla v +P_{r}(RT-\textrm{sign}(S_{0}-S_{1})\tilde{R}S)v_{3}\\ &&-(u\cdot\nabla)u\cdot v -\nabla T\cdot\nabla W +u_{3}W-(u\cdot\nabla)TW+QW-L_{e}\nabla S\cdot\nabla Z\\ &&+\textrm{sign}(S_{0}-S_{1})u_{3}Z-(u\cdot\nabla)SZ]{\rm d}x. \end{eqnarray*} $

首先, 证明条件(A1), 也就是说

$ \langle F\Phi, \Phi\rangle\leq -C_{1}\| \Phi\| _{H_{1}}^{2}+C_{2}\| \Phi\| _{H}^{2}+f(t). $

事实上, 令$\Phi=\Psi$, 则

$ \begin{eqnarray*} \langle F \Phi, \Phi\rangle&=&\int_{\Omega}[-P_{r}\nabla u\cdot\nabla u+P_{r} (RT-\textrm{sign}(S_{0}-S_{1})\tilde{R}S)u_{3}\\ &&-(u\cdot\nabla)u\cdot u-\nabla T\cdot\nabla T +u_{3}T-(u\cdot\nabla)TT+QT-L_{e}\nabla S\cdot\nabla S\\ &&+\textrm{sign}(S_{0}-S_{1})u_{3}S-(u\cdot\nabla)SS]{\rm d}x \\ &=&\int_{\Omega}[-P_{r}|\nabla u|^{2}+P_{r}(RT-\textrm{sign}(S_{0} -S_{1})\tilde{R}S)u_{3}-|\nabla T|^{2} \\ &&+Tu_{3}+QT-L_{e}|\nabla S|^{2} +\textrm{sign}(S_{0}-S_{1})u_{3}S]{\rm d}x \\ &=&\int_{\Omega}[-P_{r}|\nabla u|^{2}-|\nabla T|^{2}-L_{e}|\nabla S|^{2}+(P_{r}R+1)Tu_{3} \\ &&+(-P_{r}\tilde{R}+1)\textrm{sign}(S_{0}-S_{1})u_{3}S+QT]{\rm d}x \\ &\leq& -C_{1}\int_{\Omega}[|\nabla u|^{2}+|\nabla T|^{2}+|\nabla S|^{2}]{\rm d}x+C_{2}\int_{\Omega}[|T\| u|+|u\| S|+|Q\| T|]{\rm d}x \\ &\leq& -C_{1}\int_{\Omega}[|\nabla u|^{2}+|\nabla T|^{2}+|\nabla S|^{2}]{\rm d}x \\ &&+C_{2}\int_{\Omega}[\frac{1}{2}|T|^{2}+\frac{1}{2}|u|^{2}+\frac{1}{2}|u|^{2}+\frac{1}{2}|S|^{2} +\frac{1}{2}|T|^{2} +\frac{1}{2}|Q|^{2}]{\rm d}x \\ &\leq &-C_{1}\int_{\Omega}[|\nabla u|^{2}+|\nabla T|^{2}+|\nabla S|^{2}]{\rm d}x +C_{2}\int_{\Omega}[|u|^{2}+|T|^{2}+|S|^{2}]{\rm d}x\\ &&+C\int_{\Omega}|Q|^{2}{\rm d}x \\ &\leq& -C_{1}\| \Phi\| _{H_{1}}^{2}+C_{2}\| \Phi\| _{H}^{2}+C, \end{eqnarray*} $

即条件(A1)是满足的.

其次, 证明条件(A2).对$\forall\Phi\in C^{1}([0, \infty), X)$以及$\Psi\in X, $ $h(0 < h < 1)$, 可以得到

$ \begin{eqnarray*} &&\Big|\int^{t+h}_{t}\langle F\Phi, \Psi\rangle {\rm d}t\Big|\\ &=&\Big|\int^{t+h}_{t}\int_{\Omega}[-P_{r}\nabla u\cdot\nabla v+P_{r}(RT-\textrm{sign}(S_{0}-S_{1})\tilde{R}S)v_{3} \\ &&-(u\cdot\nabla)u\cdot v-\nabla T\cdot\nabla W+Wu_{3}-(u\cdot\nabla)TW+QW-L_{e}\nabla S\cdot\nabla Z \\ &&+\textrm{sign}(S_{0}-S_{1})u_{3}Z-(u\cdot\nabla)SZ]{\rm d}x{\rm d}t\Big| \\ &\leq& C\int^{t+h}_{t}\int_{\Omega}[|\nabla u\| \nabla v|+|T\| v|+|S\| v|+|\nabla T\| \nabla W|+|W\| u|+|Q\| W| \\ &&+|\nabla S\| \nabla Z|+|u\| Z|]{\rm d}x{\rm d}t\\ &&+C\int^{t+h}_{h} \Big[\Big|\int_{\Omega}(u\cdot\nabla)u\cdot v{\rm d}x\Big| +\Big|\int_{\Omega}(u\cdot\nabla)T\cdot W{\rm d}x\Big| +\Big|\int_{\Omega}(u\cdot\nabla)SZ{\rm d}x\Big|\Big]{\rm d}t \\ &\leq &C \Big[\int^{T}_{0}\int_{\Omega}|\nabla u|^{2}{\rm d}x{\rm d}t \Big]^{\frac{1}{2}} \Big[\int^{t+h}_{t}\int_{\Omega}|\nabla v|^{2}{\rm d}x{\rm d}t\Big]^{\frac{1}{2}} \\ && +C\Big[\int^{T}_{0}\int_{\Omega}|T|^{2}{\rm d}x{\rm d}t\Big]^{\frac{1}{2}} \Big[\int^{t+h}_{t}\int_{\Omega}|v|^{2}{\rm d}x{\rm d}t\Big]^{\frac{1}{2}} \\ &&+C\Big[\int^{T}_{0}\int_{\Omega}|S|^{2}{\rm d}x{\rm d}t\Big]^{\frac{1}{2}} \Big[\int^{t+h}_{t}\int_{\Omega}|v|^{2}{\rm d}x{\rm d}t\Big]^{\frac{1}{2}} \\ &&+C\Big[\int^{T}_{0}\int_{\Omega}|\nabla T|^{2}{\rm d}x{\rm d}t\Big]^{\frac{1}{2}} \Big[\int^{t+h}_{t}\int_{\Omega}|\nabla W|^{2}{\rm d}x{\rm d}t\Big]^{\frac{1}{2}} \\ &&+C\Big[\int^{T}_{0}\int_{\Omega}|u|^{2}{\rm d}x{\rm d}t\Big]^{\frac{1}{2}} \Big[\int^{t+h}_{t}\int_{\Omega}|W|^{2}{\rm d}x{\rm d}t\Big]^{\frac{1}{2}} \\ &&+C\Big[\int^{t+h}_{t}\int_{\Omega}|Q|^{2}{\rm d}x{\rm d}t\Big]^{\frac{1}{2}} \Big[\int^{t+h}_{t}\int_{\Omega}|W|^{2}{\rm d}x{\rm d}t\Big]^{\frac{1}{2}} \\ &&+C\Big[\int^{T}_{0}\int_{\Omega}|\nabla S|^{2}{\rm d}x{\rm d}t\Big]^{\frac{1}{2}} \Big[\int^{t+h}_{t}\int_{\Omega}|\nabla Z|^{2}{\rm d}x{\rm d}t\Big]^{\frac{1}{2}} \\ &&+C\Big[\int^{T}_{0}\int_{\Omega}|u|^{2}{\rm d}x{\rm d}t\Big]^{\frac{1}{2}} \Big[\int^{t+h}_{t}\int_{\Omega}|Z|^{2}{\rm d}x{\rm d}t\Big]^{\frac{1}{2}} \\ &&+C\int^{t+h}_{t}\int_{\Omega}\sum\limits_{i, j=1}^{3}|u_{i}|| u_{j}|\Big| \frac{\partial v_{j}}{\partial x_{i}}\Big|{\rm d}x{\rm d}t +C\int^{t+h}_{t}\int_{\Omega}\sum\limits_{i=1}^{3}|u_{i}|| T| \Big| \frac{\partial W}{\partial x_{i}}\Big|{\rm d}x{\rm d}t \\ &&+C\int^{t+h}_{t}\int_{\Omega}\sum\limits_{i=1}^{3}|u_{i}| | S| \Big| \frac{\partial Z}{\partial x_{i}}\Big|{\rm d}x{\rm d}t \\ &\leq& C\| u\| _{L^{2}((0, T), H^{1}(\Omega, \mathbb{R}^3))}\| \nabla v\| _{L^{2}(\Omega, \mathbb{R}^3)}h^{\frac{1}{2}}+C\| T\| _{L^{2}((0, T), L^{2}(\Omega))}\| v\| _{L^{2}(\Omega, \mathbb{R}^3)} h^{\frac{1}{2}} \\ &&+C\| S\| _{L^{2}((0, T), L^{2}(\Omega))}\| v\| _{L^{2}(\Omega, \mathbb{R}^3)}h^{\frac{1}{2}} +C\| T\| _{L^{2}((0, T), H^{1}(\Omega))}\| \nabla W\| _{L^{2}(\Omega)}h^{\frac{1}{2}} \\ && +C\| W\| _{L^{2}(\Omega)}\| u\| _{L^{2}((0, T), L^{2}(\Omega, \mathbb{R}^3))}h^{\frac{1}{2}} +C\| Q\| _{L^{2}(\Omega)}\| W\| _{L^{2}(\Omega)}h \\ &&+C\| S\| _{L^{2}((0, T), H^{1}(\Omega))}\| \nabla Z\| _{L^{2}(\Omega)}h^{\frac{1}{2}} +C\| u\| _{L^{2}((0, T), L^{2}(\Omega, \mathbb{R}^3))}\| Z\| _{L^{2}(\Omega)}h^{\frac{1}{2}} \\ &&+C\| v\| _{C^{1}(\Omega, \mathbb{R}^3)}\| u\| ^{2}_{L^{\infty}((0, T), L^{2}(\Omega, \mathbb{R}^3))}h \\ &&+C\| W\| _{C^{1}(\Omega)} \| u\| _{L^{\infty}((0, T), L^{2}(\Omega, \mathbb{R}^3))}\| T\| _{L^{\infty}((0, T), L^{2}(\Omega))}h \\ &&+C\| Z\| _{C^{1}(\Omega)} \| u\| _{L^{\infty}((0, T), L^{2}(\Omega, \mathbb{R}^3))}\| S\| _{L^{\infty}((0, T), L^{2}(\Omega))}h \\ &\leq& Ch^{\frac{1}{2}}, \end{eqnarray*} $

即条件(A2)满足.

最后, 证明算子$F$是T -弱连续的.设$\Phi_{n}=(u^{n}, T^{n}, S^{n})\rightharpoonup \Phi^{0}=(u^{0}, T^{0}, S^{0})$是一致收敛, 也就是说, ${\Phi_{n}}\subset L^{\infty}((0, T), H)$有界且

$ \begin{eqnarray} \label{eqc2} \left\{ \begin{array}{ll} \Phi_{n}\rightharpoonup \Phi^{0}, \ \ L^{p}((0, T), H_{1}), \\[2mm] \lim\limits_{n\rightarrow \infty}\int^{T}_{0}|\langle \Phi_{n}-\Phi^{0}, \Psi\rangle_{H}|^{2}{\rm d}t=0, \forall \Psi\in H. \end{array} \right. \end{eqnarray} $ (3.2)

根据引理2.1, 可以得到, 在$L^{2}((0, T)\times\Omega)$中, $\Phi_{n}\rightarrow \Phi^{0}$.于是, $\forall \Psi\in C^{\infty}(\Omega, \mathbb{R}^{5})\bigcap H_{1}$, 有

$ \begin{eqnarray}\label{eqc3} \lim\limits_{n\rightarrow\infty}\int^{t}_{0}\int_{\Omega}(u^{n} \cdot\nabla)u^{n}\cdot v{\rm d}x{\rm d}t &=&\lim\limits_{n\rightarrow\infty}\int^{t}_{0}\int_{\Omega}\sum\limits_{i, j=1}^{3}u^{n}_{i}\frac{\partial u^{n}_{j}}{\partial x_{i}}v_{j}{\rm d}x{\rm d}t \\ &=&-\lim\limits_{n\rightarrow\infty}\int^{t}_{0}\int_{\Omega}\sum\limits_{i, j=1}^{3} u^{n}_{i}\frac{\partial v_{j}}{\partial x_{i}}u^{n}_{j}{\rm d}x{\rm d}t\\ &=&-\int^{t}_{0}\int_{\Omega}(u^{0}\cdot\nabla)v\cdot u^{0}{\rm d}x{\rm d}t \\ &=&\int^{t}_{0}\int_{\Omega}(u^{0}\cdot\nabla)u^{0}\cdot v{\rm d}x{\rm d}t. \end{eqnarray} $ (3.3)

同理

$ \begin{equation} \label{eqc4} \lim\limits_{n\rightarrow\infty}\int^{t}_{0}\int_{\Omega}(u^{n}\cdot\nabla)T^{n}W{\rm d}x{\rm d}t =\int^{t}_{0}\int_{\Omega}(u^{0}\cdot\nabla)T^{0}W{\rm d}x{\rm d}t, \end{equation} $ (3.4)

以及

$ \begin{equation} \label{eqc5} \lim\limits_{n\rightarrow\infty}\int^{t}_{0}\int_{\Omega}(u^{n}\cdot\nabla)S^{n}Z{\rm d}x{\rm d}t =\int^{t}_{0}\int_{\Omega}(u^{0}\cdot\nabla)S^{0}Z{\rm d}x{\rm d}t. \end{equation} $ (3.5)

因此

$ \begin{equation} \label{eqc26} \lim\limits_{n\rightarrow\infty}\int^{t}_{0}\langle F\Phi_{n}, \Psi\rangle {\rm d}t=\int^{t}_{0}\langle F\Phi^{0}, \Psi\rangle {\rm d}t, \ \ \Psi\in X. \end{equation} $ (3.6)

由于$X$在$H_{1}$中稠密, 于是等式(3.6)对$\forall \Psi\in H_{1}$成立, 换句话说映射$F$是T弱连续的.因此, 根据引理2.2方程(1.1)-(1.7)存在全局弱解.

参考文献
[1] Dijkstra H, Ghil M. Low-frequency variability of the large-scale ocean circulations:a dynamical system approach. Review of Geophysics, 2005, 43(3): 1–38.
[2] Henry S, Hall P. Thermohaline convection with two stable regimes of flow. Tellus, 1961, 13(2): 224–230. DOI:10.3402/tellusa.v13i2.9491
[3] Thual O, McWilliams J. The Catastrophe structure of thermohaline convection in a two-dimensional fluid model and a comparison with low-order box models. Geophysical and Astrophysical Fluid Dynamics, 1992, 64(1-4): 67–95. DOI:10.1080/03091929208228085
[4] Tziperman E, Toggweiler J, Fzliks Y, Bryan K. Instability of the thermohaline circulation with respect to mixed boundary conditions:Is it really a problem for realistic models. Journal of Physical Oceanography, 1994, 24(2): 217–232. DOI:10.1175/1520-0485(1994)024<0217:IOTTCW>2.0.CO;2
[5] Henk A, Molemaker M. Symmetry breaking and overturning oscillations in thermohaline-driven flows. Journal of Fluid Mechanics, 1997, 331: 195–232.
[6] Dijkstra H, Neclin J. Imperfections of the thermohaline circulation:Multiple equilibria and flux correction. Journal of Climate, 1999, 12(5): 1382–1392. DOI:10.1175/1520-0442(1999)012<1382:IOTTCM>2.0.CO;2
[7] Ma T, Wang S H. Dynamic transition theory for thermohaline circulation. Physical D, 2010, 239(3/4): 167–189.
[8] Xing C, Zhou P, Luo H. The steady state solutions to the thermohaline circulation equations. Discrete Continuous Dynamical System Series B, 2016, 21(10): 3709–3722. DOI:10.3934/dcdsb
[9] 黄海洋, 郭柏灵. 复杂地形情况下大气动态模型的弱解和吸引子的存在性. 数学物理学报, 2007, 27A(6): 1098–1110.
Huang H Y, Guo B L. The existence of weak solutoins and the trajectory attractors to the model of climate dynamics. Acta Math Sci, 2007, 27A(6): 1098–1110.
[10] Alexander G R. Existence and uniqueness of the global solution to the Navier-Stokes equations. Applied Mathematics Letters, 2015, 49: 7–11. DOI:10.1016/j.aml.2015.04.008
[11] Zhang Q S, Li Y X. Global weak solutions for the three-dimensional chemotaxis-Navier-Stokes system with nonlinear diffusion. Journal of Differential Equations, 2015, 259(8): 3730–3754. DOI:10.1016/j.jde.2015.05.012
[12] Fang S M, Jin L Y, Guo B L. Existence of weak solution for quantum Zakharov equations for plasmas model. Applied Mathematics and Mechanics, 2011, 32(10): 1339–1344. DOI:10.1007/s10483-011-1504-7
[13] 马文雅, 张乔夫, 崔俊芝. 局部任意阶增长的拟线性抛物方程解的存在性. 数学物理学报, 2015, 35A(5): 833–844.
Ma W Y, Zhang Q F, Cui J Z. Global existence for some quasi-linear parabolic equations with loclly arbitrary growth and memory. Acta Math Sci, 2015, 35A(5): 833–844.
[14] Guan C X, Yin Z Y. On the existence of global weak solutions to an integrable two-component Camassa-Holm shallow-water system. Proceedings of the Edinburgh Mathematical Society, 2013, 56(3): 755–775. DOI:10.1017/S0013091513000394
[15] 马天. 偏微分方程理论与方法. 北京:科学出版社, 2011: 308–326.
Ma T. Theories and Methods in Partial Differential Equations. Peking:Science Press, 2011: 308–326.