数学物理学报, 2023, 43(2): 481-490

球对称 Chaplygin 气体相对论 Euler 方程组的奇性形成

侍迎春,, 赖耕,*

上海大学数学系 上海 200444

Formation of Singularities in Solutions to Spherically Symmetric Relativistic Euler Equations for a Chaplygin Gas

Shi Yingchun,, Lai Geng,*

Department of Mathematics, Shanghai University, Shanghai 200444

通讯作者: *赖耕,E-mail: laigeng@shu.edu.cn

收稿日期: 2022-05-21   修回日期: 2022-10-17  

基金资助: 国家自然科学基金(12071278)

Received: 2022-05-21   Revised: 2022-10-17  

Fund supported: NSFC(12071278)

作者简介 About authors

侍迎春,E-mail:shiyc@shu.edu.cn

摘要

该文研究了具有 Chaplygin 气体状态方程的相对论 Euler 方程组经典解的奇性形成. 给出了关于初值的一个充分条件, 使得 Chaplygin 气体相对论方程组的一维 Cauchy 问题经典解的质能密度 $\rho$ 本身在有限时间内发生破裂.

关键词: 相对论Euler方程组; Chaplygin 气体; 奇性形成; 特征分解

Abstract

This paper studies the formation of singularities in smooth solutions to three-dimensional (3D) spherically symmetric relativistic Euler matrixs with a Chaplygin gas matrix of state. We give a sufficient condition on the initial data to obtain that the mass-energy density itself of the classical solutions to the Cauchy problem blows up in finite time.

Keywords: Relativistic Euler matrixs; Chaplygin gas; Singularities formation; Characteristic decomposition

PDF (330KB) 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

侍迎春, 赖耕. 球对称 Chaplygin 气体相对论 Euler 方程组的奇性形成[J]. 数学物理学报, 2023, 43(2): 481-490

Shi Yingchun, Lai Geng. Formation of Singularities in Solutions to Spherically Symmetric Relativistic Euler Equations for a Chaplygin Gas[J]. Acta Mathematica Scientia, 2023, 43(2): 481-490

1 引言

在闵可夫斯基时空中相对论流体的运动由相对论 Euler 方程组来刻画, 它包含能量和动量守恒

$\begin{matrix} (T^{\alpha\beta})_{x_{\beta}}=0,\quad \alpha,\beta=0,1,2,3 \label{URE}\end{matrix}$

和粒子数守恒

$\begin{matrix}\label{par} (nu^{\alpha})_{ x_{\alpha}}=0,\end{matrix}$

其中希腊字母表示的相同的上、下标意味着按该指标从0到3求和.$T^{\alpha\beta}=iu^{\alpha}u^{\beta}-pg^{\alpha\beta}$表示能量—动量张量,$(u^{0}, u^{1}, u^{2}, u^{3})=\chi(1, u_1, u_2, u_3 )$表示流体的四维速度,$(u_1, u_2, u_3)$表示流体的宏观三维速度, $\chi=\frac{1}{\sqrt{1-u^{2}}}$ 是洛伦兹因子,$u=\sqrt{u_{1}^{2}+u_{2}^{2}+u_{3}^{2}}$ 必须小于1, 且

$g=\left( \begin{array}{cccc} 1 & 0 &0 &0 \\ 0 & -1 & 0 & 0 \\ 0 & 0 & -1 & 0 \\ 0 & 0 & 0 & -1 \\ \end{array} \right).$

其中 $n$ 表示固有粒子密度, $p$ 表示固有压强密度, $e$ 表示单位静止质量的热力学内能, $\rho=n+e$ 表示质能密度, $i=\rho+p$ 表示固有焓密度, 光速记为1[1].

本文研究三维相对论能量和动量守恒定律方程组

$\begin{matrix} \label{RE}\left\{ \begin{array}{ll} \Big(\frac{iu^2}{1-u^2}+\rho\Big)_t+\Big(\frac{iu_1}{1-u^2}\Big)_{x_1}+ \Big(\frac{iu_2}{1-u^2}\Big)_{x_2}+ \Big(\frac{iu_3}{1-u^2}\Big)_{x_3}=0,\\[3mm] \Big(\frac{iu_1}{1-u^2}\Big)_t+\Big(\frac{iu_1^2}{1-u^2}+p\Big)_{x_1}+ \Big(\frac{iu_1u_2}{1-u^2}\Big)_{x_2}+ \Big(\frac{iu_1u_3}{1-u^2}\Big)_{x_3}=0, \\[3mm] \Big(\frac{iu_2}{1-u^2}\Big)_t+\Big(\frac{iu_1u_2}{1-u^2}\Big)_{x_1}+\Big(\frac{iu_2^2}{1-u^2} +p\Big)_{x_2}+\Big(\frac{iu_2u_3}{1-u^2}\Big)_{x_3}=0, \\[3mm] \Big(\frac{iu_3}{1-u^2}\Big)_t+\Big(\frac{iu_1u_3}{1-u^2}\Big)_{x_1}+ \Big(\frac{iu_2u_3}{1-u^2}\Big)_{x_2} +\Big(\frac{iu_3^2}{1-u^2}+p\Big)_{x_3}=0. \end{array} \right.\end{matrix}$

考虑球对称流动, 即假设流动具有以下性质

$(u_1, u_2, u_3)(x_1, x_2, x_3, t)=u(x, t)\bigg(\frac{x_1}{x}, \frac{x_2}{x}, \frac{x_3}{x}\bigg),\quad \rho(x_1, x_2, x_3, t)=\rho(x, t),$

其中 $x=\sqrt{x_1^2+x_2^2+x_3^2}$. 通过该变换, 方程组 (1.3) 可化成如下形式

$\begin{matrix} \left\{ \begin{array}{ll} \Big(\frac{iu^{2}}{1-u^{2}}+\rho\Big)_t+\Big(\frac{iu}{1-u^{2}}\Big)_x+\frac{2iu}{(1-u^{2})x}=0,\\[3mm]\Big(\frac{iu}{1-u^{2}}\Big)_t+\Big(\frac{iu^{2}}{1-u^{2}}+p\Big)_x +\frac{2iu^{2}}{(1-u^{2})x}=0, \end{array} \right.\end{matrix}$

其中 $(x, t)\in (0, +\infty)\times(0, +\infty)$.

考虑如下 Chaplygin 气体状态方程

$\begin{matrix}p=-\frac{1}{\rho}. \end{matrix} $

该状态方程最早是由Chaplygin[2], 钱学森[3], 以及von Karman[4]在近似计算机翼的升力时提出的. 它还在天体物理学中被用来研究暗能量[5-8]. Brenier[9]研究了具有Chaplygin气体状态方程 (1.5) 的非相对论 Euler 方程组的一维 Riemann 问题. 他发现当 Riemann 初值满足特定条件时, 会出现质量集中现象, 即会有 $\delta$ 波出现. 最近, Cheng 和 Yang[10,11]将该结果推广到相对论情形.

近年来, 有许多关于相对论和非相对论 Chaplygin 气体 Euler 方程组的整体经典解的结果. Godin[12]得到了非等熵球对称 Chaplygin 气体 Euler 方程组在初值为一静止常状态的小扰动时的整体经典解的存在性. 关于二维轴对称 Euler 方程组相似的结果见文献[13]. 最近, Hou 和Yin[14,15]得到二维非等熵有旋轴对称 Chaplygin 气体 Euler 方程组整体经典解的存在性. Kong 等[16] 得到了二维等熵无旋 Chaplygin 气体 Euler 方程组的整体经典解. 在文献[17]中, Lei 和 Wei 将文献[12] 结果推广到三维球对称相对论 Euler 方程组.

众所周知, 无论初始数据多么小和光滑, 非相对论和相对论 Euler 方程组 Cauchy 问题的经典解都有可能在有限时间内破裂; 有关这方面的研究可以参见文献 [18-27]. 虽然 Chaplygin 气体相对论 Euler 方程组是线性退化的, 但它仍然是拟线性双曲方程组. Majda 在专著[p.89]提出了以下猜想:“如果一个双曲守恒定律方程组的所有特征都是线性退化的, 那么当初始数据光滑时, 方程组具有整体光滑解, 除非解本身在有限的时间内破裂. 特别地, 对于任何光滑的初值, 都不会形成激波.”

Qu[29]给出了一维的一阶线性退化的拟线性双曲方程组奇性形成的机理. 他使用能量方法说明解的所有范数应该是破裂的. 我们关心 Chaplygin 气体方程组的奇性形成, 特别是质量集中现象. 最近, Lai 和 Zhu[30]研究了轴对称 Chaplygin 气体 Euler 方程组的奇性形成. 在本文中, 我们把文献[30]的结果推广到球对称相对论的情形. 为此, 本文考虑方程组 (1.4) 带有如下初值

$\begin{matrix}\label{iv} (\rho,u)(x, 0)=(\rho_0, \tilde{u}(x)), \quad x\in [x_1, x_2], \end{matrix}$

其中 $x_1>0$, $\rho_0$ 是大于1的正常数, $\tilde{u}(x)\in C^1[x_1, x_2]$, $\tilde{u}(x_2)>\frac{1}{\rho_0}$, 且有

$\begin{matrix}\tilde{u}(x)>0,\quad\tilde{u}'(x)+\frac{2\tilde{u}(x)}{x}<0\quad x\in [x_1, x_2].\end{matrix}$

假设存在区间 $(a, b)\in (x_1, x_2)$, 使得如下条件成立

(A1) $\frac{u_a-c_0}{1-u_ac_0}>\frac{u_b+c_0}{1+u_bc_0}$;

(A2) $ a-x_1>\big(\frac{u_1+c_0}{1+u_1c_0}-\frac{u_a-c_0}{1-u_ac_0}\big)t_*$;

(A3) $ x_2-b>\big(\frac{u_b+c_0}{1+u_bc_0}-\frac{u_2-c_0}{1-u_2c_0}\big)t_*$;

(A4) $ m-\frac{c_0}{3}\big(\frac{1+u_bc_0}{u_b+c_0}\big)\Big\{\big(b+\frac{u_b+c_0}{1+u_bc_0}t_*\big)^3-b^3\Big\}>0$,

其中

$ c_0=\frac{1}{\rho_0}, \quad u_1=\tilde{u}(x_1), \quad u_2=\tilde{u}(x_2), \quad u_a=\tilde{u}(a), \quad u_b=\tilde{u}(b),$
$ t_*=\frac{b-a}{\frac{u_a-c_0}{1-u_ac_0}-\frac{u_b+c_0}{1+u_bc_0}}, \ \ \ m=\int_{a}^{b}x^2\cdot\frac{i_0\tilde{u}(x)}{1-\tilde{u}^2(x)}~{\rm d}x,\quad \quad i_0=\rho_0-\frac{1}{\rho_0}. $

我们将证明,在上述假设下, 问题 (1.4), (1.6) 的经典解的质能密度 $\rho$ 本身在有限时间内破裂. 主要结果如定理 2.1 所示.

注1.1 本文的目的是给出初始数据的一个充分条件来证明方程组 (1.4) 的奇性形成. 因此, 我们只需关心$[x_1, x_2]$ 上的初始数据是否会导致奇性形成.

注1.2 假设 (A1)-(A4)在某些条件下是可以满足的. 例如, 若$u_a-u_b=\frac{1}{2}$, $a=\frac{x_1+x_2}{2}$,且 $b-a$$c_0$ 是充分小的, 则满足 (A1)-(A4).

在文献[30] 中, 作者从质量守恒中得到了密度本身在有限时间内破裂的结论. 然而, 对于相对论 Euler 方程组, 质量守恒是不成立的. 我们将用动量守恒来推导质能密度 $\rho$ 本身在有限时间内破裂. 本文用特征的方法来证明奇性形成. 我们将证明通过点 $(a, 0)$$(b, 0)$ 点的轨迹线在有限时间内相交. 结合方程组(1.4)的第二个方程(动量守恒), 我们得到了质量—能量的集中.

2 球对称相对论 Euler 方程组的奇性形成

2.1 特征理论

方程组 (1.4) 可以写成如下的矩阵形式

$\begin{matrix} U_{t}+A^{-1}BU_{x}+A^{-1}C=0,\end{matrix}$

其中

$A=\left(\begin{array}{cc}\displaystyle{\frac{1+c^{2}}{1-u^{2}}}-c^{2} ~& \displaystyle{\frac{2iu}{(1-u^{2})^{2}}}\\\displaystyle{\frac{(1-u^{2})uc^{2}}{i}} ~& 1\end{array}\right),\quad B=\left(\begin{array}{cc}\displaystyle{\frac{(1+c^{2})u}{1-u^{2}}}~ &\displaystyle{\frac{i(1+u^{2})}{(1-u^{2})^{2}}} \\\displaystyle{\frac{(1-u^{2})c^{2}}{i}} ~& u\end{array}\right),$
$C=\left(\frac{2iu}{(1-u^{2})x}, ~0\right)^{T},\quad U=(\rho,~ u)^{T},$

这里的 $c=\sqrt{p'(\rho)}=\frac{1}{\rho}$ 表示声速.

矩阵 $A^{-1}B$ 的特征值是

$\begin{matrix} \lambda_{+}=\frac{u+c}{1+uc},\quad \lambda_{-}=\frac{u-c}{1-uc}.\end{matrix}$

$\lambda_{\pm}$相应的左特征向量是$l_{\pm}=\left(\frac{(1-u^2)c^2}{1-c^2}, \pm 1\right)$, 相应的右特征向量是 $r_{\pm}=\left(\frac{1-c^2}{(1-u^2)c^2}, \pm1\right)^{T}$.

通过计算可知

$\begin{matrix}\nabla_{(\rho, u)}\lambda_{\pm}\cdot r_{\pm}=0,\end{matrix}$

这表明特征值 $\lambda_{\pm}$ 是线性退化的.

将 (2.1) 左乘 $l_{\pm}$, 可得

$\begin{matrix} \left\{ \begin{array}{ll}\frac{\partial_{+}u}{1-u^{2}}-\frac{\partial_{+}c}{1-c^{2}}=-\frac{2uc}{(1+uc)x},\\[3mm]\frac{\partial_{-}u}{1-u^{2}}+\frac{\partial_{-}c}{1-c^{2}}=\frac{2uc}{(1-uc)x}. \end{array} \right.\end{matrix}$

其中 $\partial_{\pm}:=\partial_{t}+\lambda_{\pm}\partial_{x}$.

定义函数

$f(s)=\frac{1}{2}\ln\bigg(\frac{1+s}{1-s}\bigg), \quad -1<s<1.$

于是, 方程组 (2.4) 可以简化为

$\begin{matrix}\left\{ \begin{array}{ll}\displaystyle\partial_{+}f(u)-\partial_{+}f(c)=-\frac{2uc}{(1+uc)x}, \\[3mm]\displaystyle\partial_{-}f(u)+\partial_{-}f(c)=\frac{2uc}{(1-uc)x}. \end{array}\right.\end{matrix}$

2.2 特征分解

我们将使用特征分解的方法对 Cauchy 问题 (1.4), (1.6) 解的导数作先验估计. 该方法是由李杰权, 张同和郑玉玺[31]研究二维可压 Euler 方程组的拟定常简单波时提出的. 也可参阅文献[32].

命题2.1 方程组 (1.4) 中有如下交换关系成立

$\begin{matrix}\partial_{+}\partial_{-}-\partial_{-}\partial_{+}=-\frac{2(1-c^{2})}{1-u^{2}c^{2}}\cdot\frac{u}{x}(\partial_{+}-\partial_{-}).\end{matrix}$

由 (2.2) 和 (2.4)式, 可推导出

$\begin{matrix} \left\{ \begin{array}{ll} \partial_{-}\lambda_{+}=\frac{2(1-u^{2})(1-c^{2})uc}{(1-u^{2}c^{2})(1+uc)x},\\[3mm] \partial_{+}\lambda_{-}=-\frac{2(1-u^{2})(1-c^{2})uc}{(1-u^{2}c^{2})(1-uc)x}. \end{array} \right.\end{matrix}$

由 (2.2)式可知

$\begin{matrix} \partial_{x}=\frac{\partial_{+}-\partial_{-}}{\lambda_{+}-\lambda_{-}}=\frac{1-u^{2}c^{2}}{2c(1-u^{2})}(\partial_{+}-\partial_{-}). \end{matrix}$

于是有

$\begin{matrix}\partial_{+}\partial_{-}-\partial_{-}\partial_{+} &=&(\partial_{t}+\lambda_{+}\partial_{x})(\partial_{t}+\lambda_{-}\partial_{x})- (\partial_{t}+\lambda_{-}\partial_{x})(\partial_{t}+\lambda_{+}\partial_{x}) \\ &=&(\partial_{+}\lambda_{-}-\partial_{-}\lambda_{+})\partial_{x}\\ &= &-\frac{2(1-c^{2})}{1-u^{2}c^{2}}\cdot\frac{u}{x}(\partial_{+}-\partial_{-}). \end{matrix}$

证毕.

命题2.2 关于变量 $f(c)$, 有如下特征分解式

$\begin{matrix} \left\{ \begin{array}{rl} \partial_{-}\partial_+f(c) =& \frac{u(1-c^{2})}{x}\bigg[\frac{\partial_{+}f(c)}{(1-uc)^{2}}+\frac{\partial_{-}f(c)}{(1+uc)^{2}}+\frac{2\partial_{+}f(c)}{1-u^{2}c^{2}}\bigg]\\[3mm]& +\frac{c(1-u^{2})}{x}\bigg[\frac{\partial_{+}f(c)}{(1-uc)^{2}}-\frac{\partial_{-}f(c)}{(1+uc)^{2}}\bigg]-\frac{6u^{2}c}{(1-u^{2}c^{2})x^{2}},\\[3mm]\partial_{+}\partial_-f(c) =& \frac{u(1-c^{2})}{x}\bigg[\frac{\partial_{+}f(c)}{(1-uc)^{2}}+\frac{\partial_{-}f(c)}{(1+uc)^{2}}+\frac{2\partial_{-}f(c)}{1-u^{2}c^{2}}\bigg]\\[3mm]& +\frac{c(1-u^{2})}{x}\bigg[\frac{\partial_{+}f(c)}{(1-uc)^{2}}-\frac{\partial_{-}f(c)}{(1+uc)^{2}}\bigg]-\frac{6u^{2}c}{(1-u^{2}c^{2})x^{2}}. \end{array} \right.\end{matrix}$

通过计算可知

$\begin{matrix}&&\partial_+\bigg(\frac{uc}{(1-uc)x}\bigg)+\partial_-\bigg(\frac{uc}{(1+uc)x}\bigg)\\&=&\frac{\partial_+(uc)}{(1-uc)^{2}x}+\frac{\partial_-(uc)}{(1+uc)^{2}x} -\frac{uc\partial_+x}{(1-uc)x^{2}}-\frac{uc\partial_-x}{(1+uc)x^{2}}\\&=&\frac{u(1-c^{2})}{x}\bigg[\frac{\partial_{+}f(c)}{(1-uc)^{2}}+\frac{\partial_{-}f(c)}{(1+uc)^{2}}\bigg]+\frac{c(1-u^{2})}{x}\bigg[\frac{\partial_{+}f(c)}{(1-uc)^{2}}-\frac{\partial_{-}f(c)}{(1+uc)^{2}}\bigg]\\&&-\frac{2u^{2}c}{(1-u^{2}c^{2})x^{2}}\bigg[\frac{2c^{2}(1-u^{2})}{1-u^{2}c^{2}}+1\bigg]. \end{matrix}$

$f(u)$ 使用交换关系, 有

$\begin{matrix}\partial_{+}\partial_{-}f(u)-\partial_{-}\partial_{+}f(u)=-\frac{2(1-c^{2})}{1-u^{2}c^{2}}\cdot\frac{u}{x}(\partial_{+}f(u)-\partial_{-}f(u)).\end{matrix}$

结合该式和(2.5)及(2.11)式, 有

$\begin{matrix}\partial_{+}\partial_{-}f(c)+\partial_{-}\partial_{+}f(c)&=&\partial_+\bigg(\frac{2uc}{(1-uc)x}\bigg)+\partial_-\bigg(\frac{2uc}{(1+uc)x}\bigg)\\&&+\frac{2u(1-c^{2})}{(1-u^{2}c^{2})x}(\partial_{+}f(c)+\partial_{-}f(c))-\frac{8u^{2}c(1-c^{2})}{(1-u^{2}c^{2})^{2}x^{2}}\\&=&\frac{2u(1-c^{2})}{(1-u^{2}c^{2})x}(\partial_{+}f(c)+\partial_{-}f(c))+\frac{2u(1-c^{2})}{x}\bigg[\frac{\partial_{+}f(c)}{(1-uc)^{2}}+\frac{\partial_{-}f(c)}{(1+uc)^{2}}\bigg]\\&&+\frac{2c(1-u^{2})}{x}\bigg[\frac{\partial_{+}f(c)}{(1-uc)^{2}}-\frac{\partial_{-}f(c)}{(1+uc)^{2}}\bigg]-\frac{12u^{2}c}{(1-u^{2}c^{2})x^{2}}.\end{matrix}$

$f(c)$ 使用交换关系, 有

$\begin{matrix}\partial_{+}\partial_{-}f(c)-\partial_{-}\partial_{+}f(c)=-\frac{2(1-c^{2})}{1-u^{2}c^{2}}\cdot\frac{u}{x}(\partial_{+}f(c)-\partial_{-}f(c)).\end{matrix}$

于是, 结合 (2.13) 和 (2.14)式, 可得 (2.10)式.

2.3 奇性形成

Cauchy问题 (1.4), (1.6) 经典解的局部存在性可通过特征线方法得到; 见文献[33]. 也就是说, 存在一个足够小的 $T>0$, 使得问题 (1.4), (1.6) 在区域 $\Omega(T):=\{(x, t)\mid x_{+}(t)\leq x\leq x_{-}(t), 0<t<T\}$ 存在经典解 $(u,\rho)(x,t)$, 其中 $x=x_{+}(t)$ ($x=x_{-}(t)$) 表示特征曲线 $C_{+}$ ($C_{-}$). 函数$x=x_{+}(t)$$x=x_{-}(t)$分别满足

$ \left\{ \begin{array}{ll} \displaystyle \frac{{\rm d}x_{+}(t)}{{\rm d}t}=\Big(\frac{u+c}{1+uc}\Big)(x_{+}, t), & \hbox{t>0;} \\[4pt]x_{+}(0)=x_1, \end{array}\right.\quad\mbox{及}\quad \left\{ \begin{array}{ll} \displaystyle \frac{{\rm d}x_{-}(t)}{{\rm d}t}=\Big(\frac{u-c}{1-uc}\Big)(x_{-}, t), & \hbox{t>0;} \\[4pt]x_{-}(0)=x_2, \end{array}\right.$

接下来, 要对局部解作 $C^1$ 先验估计.

引理2.1 对于某个$T>0$, 若问题 (1.4), (1.6) 在区域内存在经典解. 进一步, 假设在区域 $\Omega(T)$$c>0$, 则解在区域 $\Omega(T)$ 满足

$\begin{matrix}u_2\leq u\leq u_1,\quad \partial_{+}c< 0,\quad \partial_{-}c< 0.\end{matrix}$

由(2.5)式可知

$\begin{matrix}\left\{ \begin{array}{ll}\displaystyle\partial_{t}f(u)+\lambda_{+}\partial_{x}f(u)-\partial_{t}f(c)-\lambda_{+}\partial_{x}f(c)=-\frac{2uc}{(1+uc)x}, \\[10pt]\displaystyle\partial_{t}f(u)+\lambda_{-}\partial_{x}f(u)+\partial_{t}f(c)+\lambda_{-}\partial_{x}f(c)=\frac{2uc}{(1-uc)x}. \end{array}\right.\end{matrix}$

于是, 有

$\partial_{t}f(c)=\frac{1}{2}\Big((\lambda_{+}-\lambda_{-})\partial_{x}f(u)-(\lambda_{+}+\lambda_{-})\partial_{x}f(c)\Big)+\frac{2uc}{x(1-u^2c^2)}.$

再结合(1.7)式, 可知在初始时刻有

$\begin{matrix}\partial_{\pm} f(c)= \partial_{t}f(c)=\frac{c_0}{1-c_0^2\tilde{u}^2(x)}\bigg[\tilde{u}'(x)+\frac{2\tilde{u}(x)}{x}\bigg]<0\quad \quad x\in [x_1, x_2].\end{matrix}$

$F$点是区域 $\Omega(T)$ 中的任意一点. 由 $F$ 点出发的后向 $C_{+}$ 特征曲线和 $C_{-}$ 特征曲线分别与 $x$ 轴交于 $F_{+}$ 点和 $F_{-}$ 点. 我们用表示一个由 $\widehat{F_{+}F}$, $\widehat{F_{-}F}$,及 $\widehat{F_{-}F_{+}}$ 为边界的闭区域. 下面将证明如果对于区域 $\Omega_{_F}\setminus\{F\}$ 中的任意一点都满足 $\partial_{\pm}c<0$, 那么有 $\partial_{\pm}c(F)<0$.

首先, 如果对于区域 $\Omega_{_F}\setminus\{F\}$ 中的任意一点都满足 $\partial_{\pm}c<0$, 那么通过 (2.5)式, 得到在区域$\Omega_{_F}\setminus\{F\}$ 内有 $\partial_{+}u<0$$\partial_{-}u>0$. 再与 (1.7)式结合, 可知在区域 $\Omega_{_F}$ 内有

$\begin{matrix}c(x, t)\leq c_0<u_2\leq u(x, t)\leq u_1.\end{matrix}$

假设在 $F$ 点存在 $\partial_{+}c=0$$\partial_{-}c\leq 0$. 通过 (2.10) 式的第一个方程和 (2.18)式, 可知在 $F$ 点满足

$\begin{array}{ll}\partial_{-}\partial_{+}f(c) &~=~\frac{u(1-c^2)-c(1-u^{2})}{x}\cdot\frac{\partial_{-}f(c)}{(1+uc)^{2}}-\frac{6u^{2}c}{(1-u^{2}c^{2})x^{2}}<0.\end{array}$

这就产生了矛盾.

假设在 $F$ 点存在 $\partial_{-}c=0$$\partial_{+}c\leq 0$. 通过 (2.10) 式的第二个方程, 可知在 $F$ 点满足

$\partial_{+}\partial_{-}f(c) ~=~\frac{u(1-c^2)+c(1-u^{2})}{x}\cdot\frac{\partial_{+}f(c)}{(1+uc)^{2}}-\frac{6u^{2}c}{(1-u^{2}c^{2})x^{2}}<0.$

这就产生了矛盾.

因而由连续性假设可知, 若 $\Omega_{_F}\setminus\{F\}$ 区域中的任意一点都满足 $\partial_{\pm}c<0$, 则 $\partial_{\pm}c(F)<0$ 成立.引理证毕.

从 (2.5), (2.8)式和引理 2.1 可以看出 Cauchy问题 (1.4), (1.6) 在以 $x$ 轴和 $x=x_{\pm}(t)$ ($t>0$)为边界的区域内存在整体经典解, 除非质能密度 $\rho$ 本身在有限时间内破裂.

定理2.1 存在 $t_{s}>0$ 使得问题 (1.4), (1.6) 在区域 $\Omega(t_s)$ 存在经典解, 且解满足

$\begin{matrix}\lim\limits_{t\rightarrow t_{s}^{-}}\|\rho(\cdot, t)\|_{_{0; X(t)}}=+\infty,\end{matrix}$

其中 $X(t)=\big(x_{+}(t), x_{-}(t)\big)$.

$C_{0}^a$: $x=x_{0}(t; a)$$C_{0}^b$: $x=x_{0}(t, b)$ 分别由

$\left\{ \begin{array}{ll} \displaystyle \frac{{\rm d}x_{0}(t; a)}{{\rm d}t}=u(x_{0}(t; a), t), & \hbox{t>0;} \\[4pt]x_{0}(0; a)=a, \end{array}\right.\quad\mbox{和}\quad \left\{ \begin{array}{ll} \displaystyle \frac{{\rm d}x_{0}(t; b)}{{\rm d}t}=u(x_{0}(t; b), t), & \hbox{t>0;} \\[4pt]x_{0}(0; b)=b \end{array}\right.$

来确定. 实际上 $C_{0}^a$$C_{0}^b$ 是两条分别从 $(a, 0)$ 点和 $(b, 0)$ 点出发的轨迹线.

$\partial_{0}=\partial_{t}+u\partial_{x}$. 于是有 $\partial_{0}=s_{+}\partial_{+}+s_{-}\partial_{-}$, 其中

$s_{+}=\frac{u-\lambda_{-}}{\lambda_{+}-\lambda_{-}}>0\quad s_{-}=\frac{\lambda_{+}-u}{\lambda_{+}-\lambda_{-}}>0.$

由 (2.15)式可知

$\begin{matrix}\partial_{0}c<0.\end{matrix}$

结合 (2.5) 与 (2.20)式, 有

$\begin{matrix}\partial_{0}f(u)&=&s_{+}\partial_{+}f(u)+s_{-}\partial_{-}f(u)\\&=&s_{+}\partial_{+}f(c)-s_{-}\partial_{-}f(c)-\frac{2s_{+}uc}{(1+uc)x}+\frac{2s_{-}uc}{(1-uc)x}\\&=&s_{+}\partial_{+}f(c)-s_{-}\partial_{-}f(c)=\partial_{0}f(c)-2s_{-}\partial_{-}f(c)\\&>&\partial_{0}f(c).\end{matrix}$

将该式沿着 $C_{0}^{a}$ 积分, 有

$f(u)>f(u_a)-f(c_0),$

$\ln \bigg(\frac{1+u}{1-u}\bigg)>\ln\bigg(\frac{1+u_a}{1-u_a}\bigg)-\ln\bigg(\frac{1+c_0}{1-c_0}\bigg).$

通过计算知道在 $C_{0}^{a}$ 上, 有

$\begin{matrix}u>\frac{u_a-c_0}{1-u_ac_0}.\end{matrix}$

结合 (2.5) 与 (2.20)式, 有

$\begin{matrix}\partial_{0}f(u)&=&s_{+}\partial_{+}f(u)+s_{-}\partial_{-}f(u)\\&=&s_{+}\partial_{+}f(c)-s_{-}\partial_{-}f(c)-\frac{2s_{+}uc}{(1+uc)x}+\frac{2s_{-}uc}{(1-uc)x}\\&=&s_{+}\partial_{+}f(c)-s_{-}\partial_{-}f(c)=2s_{+}\partial_{+}f(c)-\partial_{0}f(c)\\&<&-\partial_{0}f(c).\end{matrix}$

将该式沿着 $C_{0}^{b}$ 积分, 有

$\begin{matrix}u<\frac{u_b+c_0}{1+u_bc_0}.\end{matrix}$

通过计算, 再结合 (2.15)式, 有

$\begin{matrix}\partial_{+}\lambda_{+}=\frac{1-c^2}{(1+uc)^2}\partial_{+}u+ \frac{1-u^2}{(1+uc)^2}\partial_{+}c<0\end{matrix}$

$\begin{matrix}\partial_{-}\lambda_{-}=\frac{1-c^2}{(1-uc)^2}\partial_{-}u- \frac{1-u^2}{(1+uc)^2}\partial_{-}c>0.\end{matrix}$

因而, 有

$\begin{matrix}\left\{ \begin{array}{ll} \displaystyle\frac{u+c}{1+uc}<\frac{u_1+c_0}{1+u_1c_0}, & \hbox{x=x_{+}(t);} \\[10pt] \displaystyle\frac{u-c}{1-uc}>\frac{u_2-c_0}{1-u_2c_0}, & \hbox{x=x_{-}(t).} \end{array}\right.\end{matrix}$

于是, 通过假设 (A2)-(A3), 我们知道若 $x_{0}(t; a)< x_{0}(t; b)$, 则有

$\begin{matrix}x_{+}(t)<x_{0}(t; a)<x_{0}(t; b)<x_{-}(t),\end{matrix}$

这表明若曲线 $x=x_{0}(t; a)$$x=x_{0}(t; b)$ 存在, 那么它们必须在Cauchy 问题 (1.4), (1.6)的决定域内.

由方程组 (1.4) 的第二个方程可知

$\begin{matrix}&&\frac{{\rm d}}{{\rm d}t}\int_{x_{0}(t; a)}^{x_{0}(t; b)} x^2\Big(\frac{iu}{1-u^2}\Big)(x, t){\rm d}x\\&=&\int_{x_{0}(t; a)}^{x_{0}(t; b)} x^2\Big(\frac{iu}{1-u^2}\Big)_t(x, t){\rm d}x+x_0^2(t; b)\left[\Big(\frac{iu}{1-u^2}\Big)(x_0(t; b), t)\right]\frac{{\rm d}x_{0}(t; b)}{{\rm d}t}\\&&-x_0^2(t; a)\left[\Big(\frac{iu}{1-u^2}\Big)(x_0(t; a), t)\right]\frac{{\rm d}x_{0}(t; a)}{{\rm d}t}\\&=&\int_{x_{0}(t; a)}^{x_{0}(t; b)} x^2\Big(\frac{iu}{1-u^2}\Big)_t(x, t)+\Big[x^2\Big(\frac{iu^2}{1-u^2}\Big)(x, t)\Big]_x{\rm d}x\\&=&\int_{x_{0}(t; a)}^{x_{0}(t; b)} -x^2p_x(x, t) {\rm d}x\\&=&-x^2 p(x, t)\bigg|_{x_{0}(t; b)}^{x_{0}(t; a)}+\int_{x_{0}(t; a)}^{x_{0}(t; b)}2xp(x, t){\rm d}x\\&>&-c_0 x_{0}^2(t; b)>-c_0\left(b+\frac{u_b+c_0}{1+u_bc_0}t\right)^2.\end{matrix}$

因而, 由假设 (A4), 我们知道对于 $t\leq t_{*}$, 有

$\begin{matrix}\int_{x_{0}(t; a)}^{x_{0}(t; b)} x^2\Big(\frac{iu}{1-u^2}\Big)(x, t){\rm d}x>m-\frac{c_0}{3}\bigg(\frac{1+u_bc_0}{u_b+c_0}\bigg)\bigg\{\Big(b+\frac{u_b+c_0}{1+u_bc_0}t_*\Big)^3-b^3\bigg\}>0.\end{matrix}$

于是, 由假设 (A1) 和 (2.28)式可知, 若 Cauchy 问题 (1.4), (1.6) 存在整体经典解, 则存在 $0<t^*<t_*$ 使得

$x_{+}(t^*)<x_{0}(t^*; a)=x_{0}(t^*; b)<x_{-}(t^*).$

且对于 $t<t^*$, 有

$x_{+}(t)<x_{0}(t; a)<x_{0}(t; b)<x_{-}(t).$

因而, 结合 (2.15) 与 (2.28)式, 我们知道存在$t_s\in(0, t^*]$, 使得问题 (1.4), (1.6) 在区域 $\Omega(t_s)$ 内存在经典解. 此外, 有

$ \lim\limits_{t\rightarrow t_{s}^{-}}\|\rho(\cdot, t)\|_{_{0; X(t)}}=+\infty. $

定理证毕.

参考文献

Landau L D, Lifschitz E M.

Fluid Mechanics

Oxford: Pergamon, 1987

[本文引用: 1]

Chaplygin S.

On gas jets

Sci Mem Moscow Univ Math Phys, 1904, 21: 1-121

[本文引用: 1]

Tsien H S.

Two dimensional subsonic flow of compressible fluids

J Aeronaut Sci, 1939, 6: 399-407

DOI:10.2514/8.916      URL     [本文引用: 1]

Karman T V.

Compressibility effects in aerodynamics

J Aeronaut Sci, 1941, 8: 337-365

DOI:10.2514/8.10737      URL     [本文引用: 1]

Bento M C, Bertolami O, Sen A A.

Generalized Chalplygin gas, accelerated expansion and dark-energy-matter unification

Phys Rev D, 2002, 66: 043507

[本文引用: 1]

Cruz N, Lepe S, Pena F.

Dissipative generalized Chaplygin gas as phantom dark energy physics

Phys Lett B, 2007, 646: 177-182

DOI:10.1016/j.physletb.2006.12.070      URL     [本文引用: 1]

Setare M R.

Holographic Chaplygin gas model

Phys Lett B, 2007, 648: 329-332

DOI:10.1016/j.physletb.2007.03.025      URL     [本文引用: 1]

Setare M R.

Interacting holographic generalized Chaplygin gas model

Phys Lett B, 2007, 654: 1-6

DOI:10.1016/j.physletb.2007.08.038      URL     [本文引用: 1]

Brenier Y.

Solutions with concentration to the Riemann problem for one-dimensional Chaplygin gas matrixs

J Math Fluid Mech, 2005, 7: 326-331

[本文引用: 1]

Cheng H J, Yang H C.

Riemann problem for the relativistic Chaplygin Euler matrixs

J Math Anal Appl, 2011, 381: 17-26

DOI:10.1016/j.jmaa.2011.04.017      URL     [本文引用: 1]

Cheng H J, Yang H C.

Riemann problem for the isentropic relativistic Chaplygin Euler matrixs

Z Angew Math Phys, 2012, 63: 429-440

DOI:10.1007/s00033-012-0199-7      URL     [本文引用: 1]

Godin P.

Global existence of a class of smooth 3D spherically symmetric flows of Chaplygin gases with variable entropy

J Math Pures Appl, 2007, 87: 91-117

DOI:10.1016/j.matpur.2006.10.011      URL     [本文引用: 2]

Ding B B, Witt I, Yin H C.

The global smooth symmetric solution to 2-D full compressible Euler system of Chaplygin gases

J Differential Equations, 2015, 258: 445-482

DOI:10.1016/j.jde.2014.09.018      URL     [本文引用: 1]

Hou F, Yin H C.

Global smooth axisymmetric solutions to 2D compressible Euler matrixs of Chaplygin gases with non-zero vorticity

J Differential Equations, 2019, 267: 3114-3161

DOI:10.1016/j.jde.2019.03.038      URL     [本文引用: 1]

Hou F, Yin H C.

On global axisymmetric solutions to 2D compressible full Euler matrixs of Chaplygin gases

Discrete Contin Dyn Syst, 2010, 40: 1435-1492

DOI:10.3934/dcds.2020083      URL     [本文引用: 1]

Kong D X, Liu K, Wang Y.

Global existence of smooth solutions to two-dimensional compressible isentropic Euler matrixs for Chaplygin gases

Sci China Math, 2010, 53: 719-738

DOI:10.1007/s11425-010-0060-4      URL     [本文引用: 1]

Lei Z, Wei C H.

Global radial solutions to 3D relativistic Euler matrixs for non-isentropic Chaplygin gases

Math Ann, 2017, 367: 1363-1401

DOI:10.1007/s00208-016-1396-z      URL     [本文引用: 1]

Alinhac S.

Une solution approchée en grand temps des équations d'Euler compressibles axisymétriques en dimen-sion deux

Commun Partial Differ Equ, 1992, 17: 447-490

[本文引用: 1]

Alinhac S.

Temps de vie des solutions régulières des équations d'Euler compressibles axisymétriques en dimension deux

Invent Math, 1993, 111: 627-670

DOI:10.1007/BF01231301      URL     [本文引用: 1]

Athanasiou N, Zhu S G.

Formation of singularities for the relativistic Euler matrixs

J Differential Equations, 2021, 284: 284-317

DOI:10.1016/j.jde.2021.03.010      URL     [本文引用: 1]

Chen G.

Formation of singularity and smooth wave propagation for the non-isentropic compressible Euler matrixs

J Hyperbolic Differ Equ, 2011, 8: 671-690

DOI:10.1142/S0219891611002536      URL     [本文引用: 1]

We define the notion of compressive and rarefactive waves and derive the differential equations describing smooth wave steepening for the compressible Euler equations with a varying entropy profile and general pressure laws. Using these differential equations, we directly generalize Lax's singularity (shock wave) formation results (established in 1964 for hyperbolic systems with two variables) to the 3 × 3 compressible Euler equations for a polytropic ideal gas. Our results are valid globally without restriction on the size of the variation of initial data.

Chen G, Pan R H, Zhu S G.

Singularity formation for the compressible Euler matrixs

SIAM J Math Anal, 2017, 49: 2591-2614

DOI:10.1137/16M1062818      URL     [本文引用: 1]

Christodoulou D.

The formation of Shocks in 3-dimensional Fluids

European Mathematical Society, 2007

[本文引用: 1]

Christodoulou D, Miao S. Compressible Flow and Euler's Equations. Boston, Beijing: International Press; Higher Education Press, 2014

[本文引用: 1]

Guo Y, Tahvildar Z S.

Formation of singularities in relativistic fluid dynamics and in spherically symmetric plasma dynamics

Cntemp Math, 1999, 238: 151-161

[本文引用: 1]

Luk J, Speck J.

Shock formation in solutions to the 2D compressible Euler matrixs in the presence of non-zero vorticity

Invent Math, 2018, 214: 1-169

DOI:10.1007/s00222-018-0799-8      [本文引用: 1]

Pan R H, Smoller J.

Blowup of smooth solutions for relativistic Euler matrixs

Commun Math Phys, 2006, 262: 729-755

DOI:10.1007/s00220-005-1464-9      URL     [本文引用: 1]

Majda A.

Compressible fluid flow and systems of conservation laws in several space variables

Applied Mathematical Sciences, 1984, 53

Qu P.

Mechanism of singularity formation for quasilinear hyperbolic systems with linearly degenerate characteristic fields

J Differential Equations, 2011, 251: 2066-2081

DOI:10.1016/j.jde.2011.07.005      URL     [本文引用: 1]

Lai G, Zhu M.

Formation of singularities of solutions to the compressible Euler matrixs for a Chaplygin gas

Applied Mathematics Letters, 2022, 129: 107978

[本文引用: 3]

Li J Q, Zhang T, Zheng Y X.

Simple waves and a characteristic decomposition of the two dimensional compressible Euler matrixs

Comm Math Phys, 2006, 267: 1-12

DOI:10.1007/s00220-006-0033-1      URL     [本文引用: 1]

Li J Q, Zheng Y X.

Interaction of rarefaction waves of the two-dimensional self-similar Euler matrixs

Arch Ration Mech Anal, 2009, 193: 623-657

DOI:10.1007/s00205-008-0140-6      URL     [本文引用: 1]

Li T T, Yu W C.

Boundary Value Problem for Quasilinear Hyperbolic Systems

Duke University, 1985

[本文引用: 1]

/