数学物理学报, 2023, 43(5): 1417-1426

具有梯度源项和非线性边界条件的多孔介质方程组解的爆破

沈旭辉,1, 丁俊堂,2,*

1山西财经大学应用数学学院 太原 030006

2山西大学数学科学学院 太原 030006

Blow-Up Conditions of Porous Medium Systems with Gradient Source Terms and Nonlinear Boundary Conditions

Shen Xuhui,1, Ding Juntang,2,*

1School of Applied Mathematics, Shanxi University of Finance and Economics, Taiyuan 030006

2School of Mathematical Sciences, Shanxi University, Taiyuan 030006

通讯作者: * 丁俊堂,Email: djuntang@sxu.edu.cn

收稿日期: 2022-07-14   修回日期: 2023-03-23  

基金资助: 国家自然科学基金(61473180)
山西省青年科技研究基金(20210302124533)

Received: 2022-07-14   Revised: 2023-03-23  

Fund supported: NSFC(61473180)
Youth Natural Science Foundation of Shanxi Province(20210302124533)

作者简介 About authors

沈旭辉,Email:xhuishen@sxufe.edu.cn

摘要

该文主要分析下列多孔介质方程组解的爆破现象$ \left\{ \begin{array}{ll} u_{t} =\Delta u^l+f(u,v,|\nabla u|^2,t), & \\\displaystyle v_{t} =\Delta v^m+g(u,v,|\nabla v|^2,t),&x\in\Omega, \ t\in(0,t^*), \\\displaystyle \frac{\partial u}{\partial\nu}=p(u), \ \frac{\partial v}{\partial\nu}=q(v), &x\in\partial\Omega, \ t\in(0,t^*), \\\displaystyle u(x,0)=u_{0}(x), \ v(x,0)=v_{0}(x), &x\in\overline{\Omega}, \end{array} \right. $

其中 $l, m>1, \ \Omega\subset\mathbb{R}^N \ (N\geq2)$ 为具有光滑边界的有界区域. 通过使用微分不等式技术和最大值原理, 给出方程组的解在有限时刻 $t^*$ 爆破的充分条件, 并分别导出了解的爆破时刻 $t^*$ 及爆破率的上估计.

关键词: 多孔介质方程组; 爆破; 非线性边界条件

Abstract

In this paper, we consider the blow-up of solutions to the following porous medium systems:$ \left\{ \begin{array}{ll} u_{t} =\Delta u^l+f(u,v,|\nabla u|^2,t), & \\\displaystyle v_{t} =\Delta v^m+g(u,v,|\nabla v|^2,t),&x\in\Omega, \ t\in(0,t^*), \\\displaystyle \frac{\partial u}{\partial\nu}=p(u), \ \frac{\partial v}{\partial\nu}=q(v), &x\in\partial\Omega, \ t\in(0,t^*), \\\displaystyle u(x,0)=u_{0}(x), \ v(x,0)=v_{0}(x), &x\in\overline{\Omega}, \end{array} \right. $

where $l,m>1, \ \Omega\subset\mathbb{R}^N \ (N\geq2)$ is a bounded domain with smooth boundary $\partial\Omega$. Using the differential inequality techniques and the maximum principles, we give a sufficient condition to ensure that the positive solution $(u,v)$ of the above problem is a blow-up solution that blows up at a certain finite time $t^*$. An upper estimate of $t^*$ and an upper estimate of the blow-up rate of $(u,v)$ are also obtained.

Keywords: Porous medium systems; Blow-up; Nonlinear boundary condition

PDF (602KB) 元数据 多维度评价 相关文章 导出 EndNote| Ris| Bibtex  收藏本文

本文引用格式

沈旭辉, 丁俊堂. 具有梯度源项和非线性边界条件的多孔介质方程组解的爆破[J]. 数学物理学报, 2023, 43(5): 1417-1426

Shen Xuhui, Ding Juntang. Blow-Up Conditions of Porous Medium Systems with Gradient Source Terms and Nonlinear Boundary Conditions[J]. Acta Mathematica Scientia, 2023, 43(5): 1417-1426

1 引言

本文研究下列多孔介质方程组解的爆破现象

$ \begin{equation} \left\{ \begin{array}{ll} u_{t} =\Delta u^l+f(u,v,|\nabla u|^2,t), &\\\displaystyle v_{t} =\Delta v^m+g(u,v,|\nabla v|^2,t),&x\in\Omega, \ t\in(0,t^*), \\\displaystyle \frac{\partial u}{\partial\nu}=p(u), \ \frac{\partial v}{\partial\nu}=q(v), &x\in\partial\Omega, \ t\in(0,t^*), \\\displaystyle u(x,0)=u_{0}(x), \ v(x,0)=v_{0}(x), &x\in\overline{\Omega}. \end{array} \right. \end{equation} $

这里 $ l, m>1, \ \Omega\subset\mathbb{R}^N \ (N\geq2) $ 表示具有光滑边界 $\partial\Omega$ 的有界区域, $t^*$ 代表解 $(u,v)$ 的爆破时刻, $\nu$ 为边界 $\partial\Omega$ 上的外法向量. 假设 $p, q\in C^2(\mathbb{R}^+)$ 是正函数, $f, g\in C^1(\mathbb{R}^+\times \mathbb{R}^+\times\overline{\mathbb{R}^+}\times\overline{\mathbb{R}^+})$ 为非负函数, 初始值 $u_0, v_0 \in C^2(\overline{\Omega}) $ 为正函数且满足 $\partial u_0/\partial\nu=p(u_0), \ \partial v_0/\partial\nu=q(v_0), \ x\in\partial\Omega$. 根据抛物型方程组的最大值原理[[1]] 知问题 (1.1) 有正的经典解 $(u,v)$, 进一步由正则性定理 [[2]] 得 $u,v\in C^3(\Omega\times(0,t^*))\cap C^2(\overline{\Omega}\times[0,t^*))$.

在过去的几十年里, 多孔介质方程解的爆破现象受到学者们的广泛关注, 并取得了大量的研究成果 (参见文献 [3],[4],[5],[6],[7],[8],[9],[10],[11],[12],[13],[14]). 同时, 有关多孔介质方程组解的爆破研究也有丰富的结论 (参见文献 [15],[16],[17],[18],[19],[20],[21],[22],[23]). 本文的研究主要受文献 [21], [23], [24] 中研究结果的启发.

Payne-Philippin[24] 研究了如下抛物方程组解的爆破现象

$ \begin{equation} \left\{ \begin{array}{ll} u_{t} =\Delta u+d(t)f(v), \ \ v_{t} =\Delta v+k(t)g(u),&x\in\Omega, \ t\in(0,t^*), \\ \displaystyle u(x,t)=0, \ v(x,t)=0, &x\in\partial\Omega, \ t\in(0,t^*), \\ \displaystyle u(x,0)=u_{0}(x), \ \ v(x,0)=v_{0}(x), &x\in\overline{\Omega}, \end{array} \right. \end{equation} $

其中 $\Omega\subset\mathbb{R}^N \ (N\geq2)$ 表示具有光滑边界 $\partial\Omega$ 的有界区域. 基于微分不等式技术, 作者不仅得到了问题 (1.2) 的解在有限时刻爆破的充分条件以及解的爆破时刻 $t^*$ 的上界, 而且还分别给出了当 $\Omega\subset\mathbb{R}^2$$\Omega\subset\mathbb{R}^3$ 时解的爆破时刻 $t^*$ 的下界.

Xia 等[23] 分析了下列多孔介质方程组解的爆破

$ \begin{equation} \left\{ \begin{array}{ll} u_{t} =\Delta u^l+d(t)f(v), \ \ v_{t} =\Delta v^m+k(t)g(u),&x\in\Omega, \ t\in(0,t^*), \\\displaystyle u(x,t)=0, \ v(x,t)=0, &x\in\partial\Omega, \ t\in(0,t^*), \\\displaystyle u(x,0)=u_{0}(x), \ \ v(x,0)=v_{0}(x), &x\in\overline{\Omega}, \end{array} \right. \end{equation} $

其中 $l, m>1$, $\Omega\subset\mathbb{R}^N \ (N\geq2)$ 是带有光滑边界 $\partial\Omega$ 的有界区域. 使用比较原理和微分不等式技术, 他们导出了问题 (1.3) 的解的整体存在和在有限时刻爆破的充分条件. 在此基础上, 分别给出了解的爆破时刻的上界和当 $\Omega\subset\mathbb{R}^3$ 时解的爆破时刻的下界.

Shen 等[21] 考虑了下列多孔介质方程组解的爆破

$ \begin{equation} \left\{ \begin{array}{ll} u_{t} =\Delta u^l+d(t)f(v), \ \ v_{t} =\Delta v^m+k(t)g(u),&x\in\Omega, \ t\in(0,t^*), \\ \displaystyle \frac{\partial u}{\partial\nu}=p(u), \ \frac{\partial v}{\partial\nu}=q(v), &x\in\partial\Omega, \ t\in(0,t^*), \\\displaystyle u(x,0)=u_{0}(x), \ \ v(x,0)=v_{0}(x), &x\in\overline{\Omega}. \end{array} \right. \end{equation} $

这里 $l, m>1$, $\Omega\subset\mathbb{R}^N \ (N\geq2)$ 为有界区域, 且边界 $\partial\Omega$ 是光滑的. 他们通过使用微分不等式技术和 Sobolev 嵌入定理, 得到了问题 (1.4) 的解 $(u,v)$ 在有限时刻爆破的充分条件, 给出当 $\Omega\subset\mathbb{R}^N \ (N\geq2)$ 时解的爆破时刻 $t^*$ 的上界. 同时当 $\Omega\subset\mathbb{R}^N \ (N\geq3)$ 时, 他们还导出解的爆破时刻 $t^*$ 的下界.

基于文献 [21], [23], [24] 中的研究, 本文分析问题 (1.1) 解的爆破现象. 注意到问题 (1.1) 中方程的源项与问题 (1.2)-(1.4) 中方程的源项是不同的. 在问题 (1.1) 中, 源项 $f$ 包含未知函数 $u$$v$, 时间变量 $t$ 以及梯度 $|\nabla u|^2$, 源项 $g$ 包含未知函数 $u$$v$, 时间变量 $t$ 和梯度 $|\nabla v|^2$. 因此, 仅仅使用上述三篇文献中提及的微分不等式技术来研究问题 (1.1) 解的爆破是困难的. 为此, 我们将微分不等式技术与抛物型方程组解的最大值原理相结合进行研究, 从方法上保证了我们的研究工作可以顺利完成.

2 主要结论

为了得到问题 (1.1) 的解在有限时刻 $t^*$ 爆破的充分条件, 需要定义四个辅助函数

$ \begin{equation} D(x,t)=-\frac{1}{p(u(x,t))}u_t(x,t)+\theta, \ \ x\in\overline{\Omega}, \ t\in[0,t^*), \end{equation} $
$ \begin{equation} K(x,t)=-\frac{1}{q(v(x,t))}v_t(x,t)+\mu, \ \ x\in\overline{\Omega}, \ t\in[0,t^*), \end{equation} $
$ \begin{equation} E(\alpha)=\int^{+\infty}_{\alpha}\frac{1}{p(\eta)}{\rm d}\eta, \ \ \ F(\alpha)=\int^{+\infty}_{\alpha}\frac{1}{q(\eta)}{\rm d}\eta, \ \ \ \alpha\in\mathbb{R}^+, \end{equation} $

其中常数 $\theta$$\mu$

$ \begin{equation} \theta=\min\limits_{x\in\overline{\Omega}} \left\{\frac{\Delta u_0^l(x)+f(u_0(x),v_0(x),|\nabla u_0(x)|^2,0)}{p(u_0(x))}\right\}, \end{equation} $
$ \begin{equation} \mu=\min\limits_{x\in\overline{\Omega}} \left\{\frac{\Delta v_0^m(x)+g(u_0(x),v_0(x),|\nabla v_0(x)|^2,0)}{q(v_0(x))}\right\}. \end{equation} $

由 (2.3) 式容易知函数 $E$$F$ 分别存在反函数 $E^{-1}$$F^{-1}$. 本节的主要结论叙述如下.

定理2.1$(u,v)$ 为问题 (1.1) 的经典解. 假设

(i)

$ \begin{equation} \theta>0, \ \ \mu>0. \end{equation} $

${\rm (ii)}$ 对于 $r\in\mathbb{R}^+, \ s\in\mathbb{R}^+, \ h\in\overline{\mathbb{R}^+}, \ t\in\mathbb{R}^+,$

$ \begin{equation} p'(r)\geq0, \ \ \ p''(r)+(l-1)\left(\frac{p(r)}{r}\right)'\geq0, \ \ \ \left(\frac{f(r,s,h,t)}{p(r)}\right)_r-(l-1)\frac{f(r,s,h,t)}{rp(r)}\geq0, \end{equation} $
$ \begin{equation} f_s(r,s,h,t)\geq0, \ \ \ f_h(r,s,h,t)\geq0, \ \ \ f_t(r,s,h,t)\geq0. \end{equation} $

${\rm (iii)}$ 对于 $r\in\mathbb{R}^+, \ s\in\mathbb{R}^+, \ k\in\overline{\mathbb{R}^+}, \ t\in\mathbb{R}^+,$

$ \begin{equation} q'(s)\geq0, \ \ \ q''(s)+(m-1)\left(\frac{q(s)}{s}\right)'\geq0, \ \ \ \left(\frac{g(r,s,k,t)}{q(s)}\right)_s-(m-1)\frac{g(r,s,k,t)}{sq(s)}\geq0, \end{equation} $
$ \begin{equation} g_r(r,s,k,t)\geq0, \ \ \ g_k(r,s,k,t)\geq0, \ \ \ g_t(r,s,k,t)\geq0. \end{equation} $
${\rm (iv)}$
$ \begin{equation} \int_{m_1}^{+\infty}\left(\frac{1}{\theta p(\eta)} +\frac{1}{\mu q(\eta)}\right){\rm d}\eta<+\infty, \ \ m_1=\min\limits_{x\in\overline{\Omega}}\left\{u_0(x),v_0(x)\right\}, \end{equation} $

则解 $(u,v)$ 在有限时刻 $t^*$ 发生爆破且有

$t^*\leq\frac{1}{2}\int_{m_1}^{+\infty}\left(\frac{1}{\theta p(\eta)} +\frac{1}{\mu q(\eta)}\right){\rm d}\eta, $

以及

$ u(x,t)\leq E^{-1}(\theta(t^*-t)), \ \ v(x,t)\leq F^{-1}(\mu(t^*-t)), \ \ x\in\overline{\Omega}, \ t\in[0,t^*). $

第一步, 我们先证明 (2.1)-(2.2) 式定义中的函数 $D$$K$ 分别满足 $D(x,t)\leq0, $$ K(x,t)\leq0, $$ x\in\overline{\Omega}, \ t\in[0,t^*)$.

通过计算可得

$ \begin{equation} \nabla D=\frac{p'}{p^2}u_t\nabla u-\frac{1}{p}\nabla u_t, \end{equation} $
$ \begin{equation} \nabla K=\frac{q'}{q^2}v_t\nabla v-\frac{1}{q}\nabla v_t, \end{equation} $
$ \begin{equation} \Delta D=\left(\frac{p''}{p^2}-2\frac{(p')^2}{p^3}\right)u_t|\nabla u|^2+2\frac{p'}{p^2}\left(\nabla u\cdot\nabla u_t\right)+\frac{p'}{p^2}u_t\Delta u-\frac{1}{p}\Delta u_t, \end{equation} $
$ \begin{equation} \Delta K=\left(\frac{q''}{q^2}-2\frac{(q')^2}{q^3}\right)v_t|\nabla v|^2+2\frac{q'}{q^2}\left(\nabla v\cdot\nabla v_t\right)+\frac{q'}{q^2}v_t\Delta v-\frac{1}{q}\Delta v_t. \end{equation} $

假设 $a=|\nabla u|^2$ 并结合问题 (1.1), 推出

$ \begin{matrix} \nonumber D_t&=&\displaystyle\frac{p'}{p^2}\left(u_t\right)^2-\frac{1}{p}\left(u_t\right)_t =\nonumber \displaystyle\frac{p'}{p^2}\left(u_t\right)^2-\frac{1}{p}\left(\Delta u^l+f(u,v,|\nabla u|^2,t)\right)_t \\\displaystyle\nonumber &=&\displaystyle\frac{p'}{p^2}\left(u_t\right)^2-\frac{1}{p}\left(\Delta u^l+f(u,v,a,t)\right)_t \\\displaystyle\nonumber &=&\displaystyle\frac{p'}{p^2}\left(u_t\right)^2-\frac{1}{p}\left(lu^{l-1}\Delta u+(l-1)lu^{l-2}|\nabla u|^2+f\right)_t \\\displaystyle\nonumber &=&\displaystyle\frac{p'}{p^2}\left(u_t\right)^2-(l-1)lu^{l-2}\frac{1}{p}u_t\Delta u-lu^{l-1}\frac{1}{p}\Delta u_t -(l-2)(l-1)lu^{l-3}\frac{1}{p}u_t|\nabla u|^2 \\\displaystyle &&\displaystyle-2(l-1)lu^{l-2}\frac{1}{p}\left(\nabla u\cdot\nabla u_t\right)-\frac{f_u}{p}u_t-\frac{f_v}{p}v_t -2\frac{f_a}{p}\left(\nabla u\cdot\nabla u_t\right)-\frac{f_t}{p}. \end{matrix} $

使用 (2.16) 式中的推理过程, 令 $b=|\nabla v|^2$ 并结合问题 (1.1), 可得

$ \begin{matrix} \nonumber K_t&=&\displaystyle\frac{q'}{q^2}\left(v_t\right)^2-(m-1)mv^{m-2}\frac{1}{q}v_t\Delta v-mv^{m-1}\frac{1}{q}\Delta v_t\\\displaystyle\nonumber &&-(m-2)(m-1)mv^{m-3}\frac{1}{q}v_t|\nabla v|^2 -2(m-1)mv^{m-2}\frac{1}{q}\left(\nabla v\cdot\nabla v_t\right)-\frac{g_u}{q}u_t\\\displaystyle &&\displaystyle-\frac{g_v}{q}v_t -2\frac{g_b}{q}\left(\nabla v\cdot\nabla v_t\right)-\frac{g_t}{q}. \end{matrix} $

联立 (2.14) 式和 (2.16) 式, 有

$ \begin{matrix} lu^{l-1}\Delta D-D_t \nonumber&=&\displaystyle\left[lu^{l-1}\left(\frac{p''}{p^2}-2\frac{(p')^2}{p^3}\right) +(l-2)(l-1)lu^{l-3}\frac{1}{p}\right]u_t|\nabla u|^2 \\\displaystyle \nonumber&&\displaystyle+\left(2lu^{l-1}\frac{p'}{p^2}+2(l-1)lu^{l-2}\frac{1}{p}+2\frac{f_a}{p}\right)\left(\nabla u\cdot\nabla u_t\right) \\\displaystyle &&\displaystyle+\left(lu^{l-1}\frac{p'}{p^2}+(l-1)lu^{l-2}\frac{1}{p}\right)u_t\Delta u -\frac{p'}{p^2}\left(u_t\right)^2 +\frac{f_u}{p}u_t+\frac{f_v}{p}v_t+\frac{f_t}{p}.\qquad \end{matrix} $

同样地, 使用 (2.15) 式, (2.17) 式和在 (2.18) 式中的类似计算可推出

$ \begin{matrix} mv^{m-1}\Delta K-K_t \displaystyle \nonumber&=&\displaystyle\left[mv^{m-1}\left(\frac{q''}{q^2}-2\frac{(q')^2}{q^3}\right) +(m-2)(m-1)mv^{m-3}\frac{1}{q}\right]v_t|\nabla v|^2 \\\displaystyle \nonumber&&\displaystyle+\left(2mv^{m-1}\frac{q'}{q^2}+2(m-1)mv^{m-2}\frac{1}{q}+2\frac{g_b}{q}\right)\left(\nabla v\cdot\nabla v_t\right) \\\displaystyle \nonumber&&\displaystyle+\left(mv^{m-1}\frac{q'}{q^2}+(m-1)mv^{m-2}\frac{1}{q}\right)v_t\Delta v -\frac{q'}{q^2}\left(v_t\right)^2 +\frac{g_u}{q}u_t \\\displaystyle &&\displaystyle+\frac{g_v}{q}v_t+\frac{g_t}{q}. \end{matrix} $

由 (2.12) 式和 (2.13) 式, 我们有

$ \begin{equation} \nabla u_t=-p\nabla D+\frac{p'}{p}u_t\nabla u \end{equation} $

$ \begin{equation} \nabla v_t=-q\nabla K+\frac{q'}{q}v_t\nabla v. \end{equation} $

将 (2.20) 式代入 (2.18) 式, 可得

$ \begin{matrix} \nonumber&&\displaystyle lu^{l-1}\Delta D+\frac{2l}{p}\left[\left(u^{l-1}p\right)'+\frac{pf_a}{l}\right]\left(\nabla u\cdot\nabla D\right)-D_t \\\displaystyle \nonumber&=&\displaystyle\left(lu^{l-1}\frac{p''}{p^2}+2(l-1)lu^{l-2}\frac{p'}{p^2}+(l-2)(l-1)lu^{l-3}\frac{1}{p} +2\frac{p'f_a}{p^2}\right)u_t|\nabla u|^2 \\\displaystyle &&\displaystyle+\left(lu^{l-1}\frac{p'}{p^2}+(l-1)lu^{l-2}\frac{1}{p}\right)u_t\Delta u-\frac{p'}{p^2}\left(u_t\right)^2 +\frac{f_u}{p}u_t+\frac{f_v}{p}v_t+\frac{f_t}{p}. \end{matrix} $

类似地, 将 (2.21) 式代入 (2.19) 式可以推出

$ \begin{matrix} \nonumber&&\displaystyle mv^{m-1}\Delta K+\frac{2m}{q}\left[\left(v^{m-1}q\right)'+\frac{qg_{b}}{m}\right]\left(\nabla v\cdot\nabla K\right)-K_t \\\displaystyle \nonumber&=&\displaystyle\left(mv^{m-1}\frac{q''}{q^2}+2(m-1)mv^{m-2}\frac{q'}{q^2}+(m-2)(m-1)mv^{m-3}\frac{1}{q} +2\frac{q'g_{b}}{q^2}\right)v_t|\nabla v|^2 \\\displaystyle &&\displaystyle+\left(mv^{m-1}\frac{q'}{q^2}+(m-1)mv^{m-2}\frac{1}{q}\right)v_t\Delta v-\frac{q'}{q^2}\left(v_t\right)^2 +\frac{g_u}{q}u_t+\frac{g_v}{q}v_t+\frac{g_t}{q}. \end{matrix} $

由问题 (1.1) 知

$ \begin{equation} \Delta u=\frac{1}{lu^{l-1}}u_t-\frac{l-1}{u}|\nabla u|^2-\frac{f}{lu^{l-1}}, \end{equation} $
$ \begin{equation} \Delta v=\frac{1}{mv^{m-1}}v_t-\frac{m-1}{v}|\nabla v|^2-\frac{g}{mv^{m-1}}. \end{equation} $

将 (2.24) 式入 (2.22) 式, 可得

$ \begin{matrix} \nonumber&&\displaystyle lu^{l-1}\Delta D+\frac{2l}{p}\left[\left(u^{l-1}p\right)'+\frac{pf_a}{l}\right]\left(\nabla u\cdot\nabla D\right)-D_t \\ \displaystyle \nonumber&=&\displaystyle\left(lu^{l-1}\frac{p''}{p^2}+(l-1)lu^{l-2}\frac{p'}{p^2}-(l-1)lu^{l-3}\frac{1}{p} +2\frac{p'f_a}{p^2}\right)u_t|\nabla u|^2 \\\displaystyle &&\displaystyle+\frac{l-1}{up}\left(u_t\right)^2 +\left(\frac{f_u}{p}-\frac{fp'}{p^2}-(l-1)\frac{f}{up}\right)u_t +\frac{f_v}{p}v_t+\frac{f_t}{p}. \end{matrix} $

同时将 (2.25) 式代入 (2.23) 式, 有

$\begin{matrix} \nonumber&&\displaystyle mv^{m-1}\Delta K+\frac{2m}{q}\left[\left(v^{m-1}q\right)'+\frac{qg_{b}}{m}\right]\left(\nabla v\cdot\nabla K\right)-K_t \\\displaystyle \nonumber&=&\displaystyle\left(mv^{m-1}\frac{q''}{q^2}+(m-1)mv^{m-2}\frac{q'}{q^2}-(m-1)mv^{m-3}\frac{1}{q} +2\frac{q'g_{b}}{q^2}\right)v_t|\nabla v|^2 \\\displaystyle &&\displaystyle+\frac{m-1}{vq}\left(v_t\right)^2 +\left(\frac{g_v}{q}-\frac{gq'}{q^2}-(m-1)\frac{g}{vq}\right)v_t +\frac{g_u}{q}u_t+\frac{g_t}{q}. \end{matrix}$

根据 (2.1) 式和 (2.2) 式可知

$ \begin{equation} u_t=-pD+\theta p, \ \ v_t=-qK+\mu q. \end{equation} $

结合 (2.26) 和 (2.28) 式, 可得

$ \begin{matrix} \nonumber&&\displaystyle lu^{l-1}\Delta D+\frac{2l}{p}\left[\left(u^{l-1}p\right)'+\frac{pf_a}{l}\right]\left(\nabla u\cdot\nabla D\right) \\\displaystyle \nonumber&&\displaystyle+\left\{\frac{1}{p}\left[lu^{l-1}\left(p''+(l-1)\left(\frac{p}{u}\right)'\right)\!+\!2p'f_a\right]|\nabla u|^2 +p\left[\left(\frac{f}{p}\right)_u\!-\!(l-1)\frac{f}{up}\right]\right\}D+\frac{qf_v}{p}K-D_t \\\displaystyle \nonumber&=&\displaystyle\frac{\theta}{p}\left\{lu^{l-1}\left[p''+(l-1)\left(\frac{p}{u}\right)'\right]+2p'f_a\right\}|\nabla u|^2 +\frac{l-1}{up}\left(u_t\right)^2\\\displaystyle &&\displaystyle+\theta p\left[\left(\frac{f}{p}\right)_u-(l-1)\frac{f}{up}\right] +\mu\frac{qf_v}{p}+\frac{f_t}{p}. \end{matrix} $

将 (2.28) 式代入 (2.27) 式, 我们有

$ \begin{matrix} \nonumber&&\displaystyle mv^{m-1}\Delta K+\frac{2m}{q}\left[\left(v^{m-1}q\right)'+\frac{qg_b}{m}\right]\left(\nabla v\cdot\nabla K\right)+\frac{pg_u}{q}D \\\displaystyle \nonumber&&\displaystyle+\left\{\frac{1}{q}\left[mv^{m-1}\left(q''+(m-1)\left(\frac{q}{v}\right)'\right)+2q'g_b\right]|\nabla v|^2+q\left[\left(\frac{g}{q}\right)_v-(m-1)\frac{g}{vq}\right]\right\}K-K_t \\\displaystyle \nonumber&=&\displaystyle\frac{\mu}{q}\left\{mv^{m-1}\left[q''+(m-1)\left(\frac{q}{v}\right)'\right]+2q'g_b\right\}|\nabla v|^2 +\frac{m-1}{vq}\left(v_t\right)^2 \\\displaystyle &&\displaystyle+\mu q\left[\left(\frac{g}{q}\right)_v-(m-1)\frac{g}{vq}\right] +\theta\frac{pg_u}{q}+\frac{g_t}{q}. \end{matrix} $

利用 (2.7) 式和 (2.8) 式知 (2.29) 式的右端项为非负的. 同理由 (2.9) 式和 (2.10) 式知 (2.30) 式的右端项也是非负的. 因此,

$ \begin{matrix} \nonumber&&\displaystyle lu^{l-1}\Delta D+\frac{2l}{p}\left[\left(u^{l-1}p\right)'+\frac{pf_a}{l}\right]\left(\nabla u\cdot\nabla D\right) \\\displaystyle \nonumber&&\displaystyle+\left\{\frac{1}{p}\left[lu^{l-1}\left(p''+(l-1)\left(\frac{p}{u}\right)'\right)+2p'f_a\right]|\nabla u|^2 +p\left[\left(\frac{f}{p}\right)_u-(l-1)\frac{f}{up}\right]\right\}D \\\displaystyle &&\displaystyle+\frac{qf_v}{p}K-D_t\geq0, \ \ x\in\Omega \ \ t\in(0,t^*), \end{matrix} $
$ \begin{matrix} \nonumber&&\displaystyle mv^{m-1}\Delta K+\frac{2m}{q}\left[\left(v^{m-1}q\right)'+\frac{qg_b}{m}\right]\left(\nabla v\cdot\nabla K\right)+\frac{pg_u}{q}D \\\displaystyle \nonumber&&\displaystyle+\left\{\frac{1}{q}\left[mv^{m-1}\left(q''+(m-1)\left(\frac{q}{v}\right)'\right) +2q'g_b\right]|\nabla v|^2\right. \\\displaystyle &&\displaystyle\left.+q\left[\left(\frac{g}{q}\right)_v-(m-1)\frac{g}{vq}\right]\right\}K -K_t\geq0, \ \ x\in\Omega, \ \ t\in(0,t^*). \end{matrix} $

使用问题 (1.1) 的边界条件, 可得

$ \begin{matrix} \nonumber\displaystyle \frac{\partial D}{\partial\nu}&=&\displaystyle\frac{p'u_t}{p^2}\frac{\partial u}{\partial\nu}-\frac{1}{p}\frac{\partial u_t}{\partial\nu} =\frac{p'u_t}{p}-\frac{1}{p}\left(\frac{\partial u}{\partial\nu}\right)_t =\frac{p'u_t}{p}-\frac{1}{p}p_t \\ \displaystyle &=&\displaystyle\frac{p'u_t}{p}-\frac{p'u_t}{p}=0, \ \ x\in\partial\Omega, \ \ t\in(0,t^*). \end{matrix} $

同理, 再次应用问题 (1.1) 的边界条件并重复 (2.33) 式中的计算步骤, 还可得到

$ \begin{equation} \frac{\partial K}{\partial\nu}=0, \ \ x\in\Omega, \ \ t\in(0,t^*). \end{equation} $

又根据 (2.4) 式中的 $\theta$

$ \begin{equation} D(x,0)=\theta-\frac{\Delta u_{0}^{l}(x)+f\left(u_0(x),v_0(x),|\nabla u_0(x)|^2,0\right)}{p(u_0(x))} \leq0, \ \ x\in\overline{\Omega}. \end{equation} $

类似地, 由 (2.5) 式中的常数 $\mu$ 得到

$ \begin{equation} K(x,0)\leq0, \ \ x\in\overline{\Omega}. \end{equation} $

结合 (2.31)-(2.36) 式与抛物方程组的最大值原理[[1]], 则有

$ \begin{equation} D(x,t)\leq0, \ K(x,t)\leq0, \ \ x\in\overline{\Omega}, \ \ t\in[0,t^*). \end{equation} $

第二步, 我们将使用微分不等式技术和反证法证明问题的解 $(u,v)$ 在有限时刻 $t^*$ 发生爆破. 为此, 我们假设 $(u,v)$ 是整体存在的, 那么有

$ \begin{equation} 0\leq u(x,t)<+\infty, \ \ 0\leq v(x,t)<+\infty, \ \ x\in\overline{\Omega}, \ \ t\in[0,+\infty). \end{equation} $

由 (2.37) 式可知

$ \begin{equation} \frac{1}{\theta p(u)}u_t\geq1, \ \ \frac{1}{\mu q(v)}v_t\geq1. \end{equation} $

根据 (2.39) 式可得

$ \begin{equation} \frac{1}{\theta p(u)}u_t+ \frac{1}{\mu q(u)}v_t\geq2. \end{equation} $

对于 (2.40) 式两端从 $0$$t$ 进行积分, 有

$ \begin{equation} \int^t_0\left(\frac{u_t}{\theta p(u)}+\frac{v_t}{\mu q(v)}\right){\rm d}t =\int^{u(x,t)}_{u_0(x)}\frac{1}{\theta p(\eta)}{\rm d}\eta+\int^{v(x,t)}_{v_0(x)}\frac{1}{\mu q(\eta)} {\rm d}\eta\geq 2t, \end{equation} $

其中 $x\in\overline{\Omega}$ 是一个固定点. 结合 (2.38) 式和 (2.41) 式, 我们导出

$ \begin{matrix} \nonumber\displaystyle\int_{m_1}^{+\infty}\left(\frac{1}{\theta p(\eta)} +\frac{1}{\mu q(\eta)}\right){\rm d}\eta &=&\int^{+\infty}_{m_1}\frac{1}{\theta p(\eta)}{\rm d}\eta +\int^{+\infty}_{m_1}\frac{1}{\mu q(\eta)}{\rm d}\eta \\\displaystyle &>&\displaystyle\int^{u(x,t)}_{u_0(x)}\frac{1}{\theta p(\eta)}{\rm d}\eta+\int^{v(x,t)}_{v_0(x)}\frac{1}{\mu q(\eta)}{\rm d}\eta\geq 2t. \end{matrix} $

$t\rightarrow+\infty$ 时, 则从 (2.42) 式得到

$\int_{m_1}^{+\infty}\left(\frac{1}{\theta p(\eta)} +\frac{1}{\mu q(\eta)}\right){\rm d}\eta=+\infty.$

这与 (2.11) 式相矛盾, 因此证得解 $(u,v)$ 在有限时刻 $t^*$ 发生爆破.

第三步, 给出解 $(u,v)$ 的爆破时刻 $t^*$ 及爆破率的上估计. 事实上, 由解 $(u,v)$ 在时刻 $t^*$ 爆破可知

$ \lim\limits_{t\rightarrow t^{*-}}u(x,t)=+\infty \ \ 或 \ \lim\limits_{t\rightarrow t^{*-}}v(x,t)=+\infty. $

因此, 在 (2.41) 式中令 $t\rightarrow t^{*-}$, 可得

$\begin{eqnarray*}\nonumber t^*&\leq&\displaystyle\frac{1}{2}\left(\lim\limits_{t\rightarrow t^{*-}}\int^{u(x,t)}_{u_0(x)}\frac{1}{\theta p(\eta)}{\rm d}\eta +\lim\limits_{t\rightarrow t^{*-}}\int^{v(x,t)}_{v_0(x)}\frac{1}{\mu q(\eta)}{\rm d}\eta\right)\\\displaystyle &\leq&\displaystyle\frac{1}{2}\left(\int^{+\infty}_{u_0(x)} \frac{1}{\theta p(\eta)}{\rm d}\eta +\int^{+\infty}_{v_0(x)} \frac{1}{\mu q(\eta)}{\rm d}\eta\right)\\\displaystyle &\leq&\displaystyle\frac{1}{2}\left(\int^{+\infty}_{m_1} \frac{1}{\theta p(\eta)}{\rm d}\eta +\int^{+\infty}_{m_1} \frac{1}{\mu q(\eta)}{\rm d}\eta\right)\\\displaystyle &=&\displaystyle\frac{1}{2}\int_{m_1}^{+\infty}\left(\frac{1}{\theta p(\eta)} +\frac{1}{\mu q(\eta)}\right){\rm d}\eta. \end{eqnarray*}$

此外, 对于 (2.39) 式中的关于 $u$ 的不等式从 $t$$t_0$ 进行积分可推出

$ \begin{matrix} \nonumber E(u(x,t))&\geq&\displaystyle E(u(x,t))-E(u(x,t_0)) =\int^{+\infty}_{u(x,t)}\frac{1}{p(\eta)}{\rm d}\eta -\int^{+\infty}_{u(x,t_0)}\frac{1}{p(\eta)}{\rm d}\eta \\\displaystyle &=&\int^{u(x,t_0)}_{u(x,t)}\frac{1}{p(\eta)} {\rm d}\eta =\displaystyle\int^{t_0}_{t}\frac{u_t}{p(u)}{\rm d}t \geq\theta(t_0-t), \end{matrix} $

其中 $0\leq t<t_0<t^*$$x\in\overline{\Omega}$ 是一个固定点. 在 (2.43) 式中令 $t_0\rightarrow t^{*-}$

$ \begin{equation} E(u(x,t))\geq\theta(t^*-t). \end{equation} $

同时注意到

$E'(\alpha)=-\frac{1}{p(\alpha)}<0, \ \alpha\in\mathbb{R}^{+}.$

这意味着 $E$$\mathbb{R}^{+}$ 上是严格的减函数. 因此由 (2.44) 式推出

$ \begin{equation} u(x,t)\leq E^{-1}(\theta(t^*-t)), \ \ x\in\overline{\Omega}, \ \ t\in[0,t^*). \end{equation} $

类似地, 对 (2.39) 式中关于 $v$ 的不等式重复 (2.43)-(2.45) 式中的计算过程, 我们可得

$v(x,t)\leq F^{-1}(\mu(t^*-t)), \ \ x\in\overline{\Omega}, \ \ t\in[0,t^*).$

证毕.

3 应用

本节, 将通过下列例子来说明定理 2.1 的结论.

例3.1$(u,v)$ 为下列多孔介质方程组的经典解

$\begin{equation*} \left\{ \begin{array}{ll} u_t=\Delta u^2+\left(1+tv|\nabla u|^2\right)u^4, \ \ v_t=\Delta v^2+\left(1+tu|\nabla v|^2\right)v^4, &x\in\Omega, \ t\in(0,t^*), \\\displaystyle \frac{\partial u}{\partial\nu}=\frac{1}{2}u^2, \ \ \frac{\partial v}{\partial\nu}=\frac{1}{2}v^2, &x\in\partial\Omega, \ t\in(0,t^*), \\\displaystyle u(x,0)=x^2_1+x^2_2+x^2_3+1, \ \ v(x,0)=x^2_1+x^2_2+x^2_3+1, &x\in\overline{\Omega}, \end{array} \right. \end{equation*}$

其中 $\Omega=\left\{x=(x_{1},x_{2},x_{3}) \ \left| \ x^2_1+x^2_2+x^2_3<1\right\}\right.$. 这里我们有

$ \begin{array}{c} \displaystyle l=m=2, \ \ p(u)=\frac{1}{2}u^2, \ \ q(v)=\frac{1}{2}v^2, \ \ f(u,v,|\nabla u|^2,t)=\left(1+tv|\nabla u|^2\right)u^4, \\\displaystyle g(u,v,|\nabla v|^2,t)=\left(1+tu|\nabla v|^2\right)v^4, \ \ u_0(x)=v_0(x)=x^2_1+x^2_2+x^2_3+1. \end{array} $

结合 (2.4) 式和 (2.5) 式, 有

$ \theta=\min\limits_{x\in\overline{\Omega}} \left\{\frac{\Delta u_0^l+f\left(u_0(x),v_0(x),|\nabla u_0(x)|^2,0\right)}{p(u_0(x))}\right\} =\min\limits_{u_0\in [1,2]} \left\{\frac{40}{u_{0}}-\frac{16}{u_0^2}+2u_0^2\right\} =23.7572 $

$ \mu=\min\limits_{x\in\overline{\Omega}} \left\{\frac{\Delta v_0^m+g\left(u_0(x),v_0(x),|\nabla v_0(x)|^2,0\right)}{q(v_0(x))}\right\} =\min\limits_{v_0\in [1,2]} \left\{\frac{40}{v_{0}}-\frac{16}{v_0^2}+2v_0^2\right\} =23.7572, $

这就意味着条件 (2.6) 成立. 进一步, 通过计算容易验证 (2.7)-(2.11) 式成立. 因此, 由定理 2.1 可知 $(u,v)$ 在有限时刻 $t^*$ 爆破且爆破时刻 $t^*$ 满足

$ t^*\leq\frac{1}{2}\int_{m_1}^{+\infty}\left(\frac{1}{\theta p(\eta)} +\frac{1}{\mu q(\eta)}\right){\rm d}\eta =0.0842\int^{+\infty}_{1}\frac{1}{\eta^2}{\rm d}\eta=0.0842. $

由于

$ E(\alpha)=\int^{+\infty}_{\alpha}\frac{1}{p(\eta)}{\rm d}\eta=2\int^{+\infty}_{\alpha}\frac{1}{\eta^2}{\rm d}\eta =\frac{2}{\alpha}, \ \ \alpha\in {\mathbb{R}^+}, $
$ F(\alpha)=\int^{+\infty}_{\alpha}\frac{1}{q(\eta)}{\rm d}\eta=2\int^{+\infty}_{\alpha}\frac{1}{\eta^2}{\rm d}\eta =\frac{2}{\alpha}, \ \ \alpha\in {\mathbb{R}^+}, $

我们得解的爆破率上估计

$ u(x,t)\leq E^{-1}\left(\theta(t^*-t)\right)=\frac{0.0842}{t^*-t}, \ \ x\in\overline{\Omega}, \ t\in(0,t^*), $
$ v(x,t)\leq F^{-1}\left(\mu(t^*-t)\right)=\frac{0.0842}{t^*-t}, \ \ x\in\overline{\Omega}, \ t\in(0,t^*). $

参考文献

Protter M H, Weinberger H F. Maximum Principles in Differential Equations. Englewood Cliffs: Prentice-Hall, 1967

[本文引用: 2]

Friedman A. Partial Differential Equation of Parabolic Type. Englewood Cliffs: Prentice-Hall, 1964

[本文引用: 1]

Akagi G, Melchionna S.

Porous medium equation with a blow-up nonlinearity and a non-decreasing constraint

Nonlinear Differ Equ Appl, 2019, 26(2): Article 10

[本文引用: 1]

Anderson J R, Deng K.

Global solvability for the porous medium equation with boundary flux governed by nonlinear memory

J Math Anal Appl, 2015, 423(2): 1183-1202

DOI:10.1016/j.jmaa.2014.10.041      URL     [本文引用: 1]

Andreu F, Mazón J M, Toledo J, Rossi J D.

Porous medium equation with absorption and a nonlinear boundary condition

Nonlinear Anal TMA, 2002, 49(4): 541-563

DOI:10.1016/S0362-546X(01)00122-5      URL     [本文引用: 1]

Ding J T, Shen X H.

Blow-up time estimates in porous medium equations with nonlinear boundary conditions

Z Angew Math Phys, 2018, 69(4): Article 99

[本文引用: 1]

Iagar R G, Sánchez A,

Blow up profiles for a quasilinear reaction-diffusion equation with weighted reaction

J Differential Equations, 2021, 272: 560-605

DOI:10.1016/j.jde.2020.10.006      URL     [本文引用: 1]

Li F C, Xie C H.

Global existence and blow-up for a nonlinear porous medium equation

Appl Math Lett, 2003, 16(2): 185-192

DOI:10.1016/S0893-9659(03)80030-7      URL     [本文引用: 1]

Liang Z L.

Blow up rate for a porous medium equation with power nonlinearity

Nonlinear Anal TMA, 2010, 73(11): 3507-3512

DOI:10.1016/j.na.2010.06.078      URL     [本文引用: 1]

Schaefer P W.

Blow-up phenomena in some porous medium problems

Dynam Systems Appl, 2009, 18(1): 103-109

[本文引用: 1]

Tian H M, Zhang L L.

Blow-up solution of a porous medium equation with nonlocal boundary conditions

Complexity, 2020, 2020: Article ID 9037287

[本文引用: 1]

Wang Z Y, Yin J X.

Note on blow-up of solutions for a porous medium equation with convection and boundary flux

Colloq Math, 2012, 128(2): 223-228

DOI:10.4064/cm128-2-7      URL     [本文引用: 1]

Wu X L, Gao W J.

Blow-up of the solution for a class of porous medium equation with positive initial energy

Acta Math Sci, 2013, 33B(4): 1024-1030

[本文引用: 1]

Zhou J.

A multi-dimension blow-up problem to a porous medium diffusion equation with special medium void

Appl Math Lett, 2014, 30: 6-11

DOI:10.1016/j.aml.2013.12.003      URL     [本文引用: 1]

Du L L.

Blow-up for a degenerate reaction-diffusion system with nonlinear localized sources

J Math Anal Appl, 2006, 324(1): 304-320

DOI:10.1016/j.jmaa.2005.11.052      URL     [本文引用: 1]

Lei P D, Zheng S N.

Global and nonglobal weak solutions to a degenerate parabolic system

J Math Anal Appl, 2006, 324(1): 177-198

DOI:10.1016/j.jmaa.2005.12.012      URL     [本文引用: 1]

Li Y X, Gao W J, Han Y Z.

Boundedness of global solutions for a porous medium system with moving localized sources

Nonlinear Anal TMA, 2010, 72(6): 3080-3090

DOI:10.1016/j.na.2009.11.047      URL     [本文引用: 1]

Mu C L, Hu X G, Li Y H, Cui Z J.

Blow-up and global existence for a coupled system of degenerate parabolic equations in a bounded domain

Acta Math Sci, 2007, 27B(1): 92-106

[本文引用: 1]

Mu C L, Su Y.

Global existence and blow-up for a quasilinear degenerate parabolic system in a cylinder

Appl Math Lett, 2001, 14(6): 715-723

DOI:10.1016/S0893-9659(01)80032-X      URL     [本文引用: 1]

Quir${\rm \acute{o}}$s J D, Rossi J D.

Blow-up sets and Fujita type curves for a degenerate parabolic system with nonlinear boundary conditions

Indiana Univ Math J, 2001, 50(1): 629-654

DOI:10.1512/iumj.2001.50.1828      URL     [本文引用: 1]

Shen X H, Ding J T.

Blow-up phenomena in porous medium equation systems with nonlinear boundary conditions

Comput Math Appl, 2019, 77(12): 3250-3263

DOI:10.1016/j.camwa.2019.02.007      URL     [本文引用: 4]

Wang Z J, Zhou Q A, Lou W Q.

Critical exponents for porous medium systems coupled via nonlinear boundary flux

Nonlinear Anal TMA, 2009, 71(5/6): 2134-2140

DOI:10.1016/j.na.2009.01.047      URL     [本文引用: 1]

Xia A Y, Fan M S, Li S.

Blow-up and life span estimates for a class of nonlinear degenerate parabolic system with time-dependent coefficients

Acta Math Sci, 2017, 37B(4): 974-984

[本文引用: 4]

Payne L E, Philippin G A.

Blow-up phenomena for a class of parabolic systems with time dependent coefficients

Appl Math, 2012, 3(4): 325-330

[本文引用: 3]

/