Spreading of the free boundary of relativistic Euler equations in a vacuum (Q2424332)

Let \(S_T=\{(t,x): x\in\Omega_t, \, 0<t<T\}\) be bounded domain with \(C^1\) boundary \(\mathrm{B}_T = \{(t,x): x\in\partial\Omega_t, \, 0<t<T\},\) \(x\) is \(n\)-dimensional vector. There is considered the following problem with unknown functions \((\rho,v^1,\ldots,v^n)\) and free boundary \(\partial\Omega_t\): \[\left\{ \begin{array}{c}\partial_t\hat{\rho} + \nabla_x({\tilde{\rho}}v)=0 \ \text{ in } \ S_T, \\ \partial_t({\tilde{\rho}}v) + \nabla_x({\tilde{\rho}}v \otimes v) + \nabla_x p =0 \ \text{ in } \ S_T, \end{array} \right.\ \eqno(1)\] \[\rho = 0 \ \text{on} \ \text{B}_T, \eqno(2)\] where \[ \hat{\rho} = \frac{\rho c^2 + p}{c^2 - v^2} - \frac{ p}{c^2}, \ \tilde{\rho} = \frac{\rho c^2 + p}{c^2 - v^2}, \ p = \sigma^2\rho. \] Assuming that the problem (1), (2) has solution \(\rho \in C^0(\bar{S}_T)\cap C^1(S_T), \ v \in C^1(\bar{S}_T)\) and under some conditions on the coefficients and functions the authors prove that \([\mathrm{diam} \, \Omega_t]^2 \ge k_1t^2 + k_2 t +k_3\) and \([\mathrm{diam} \Omega_t]^2 \ge k_4t^2 + k_2 t +k_3\) for \(0\le t <T,\) in the cases \(\sigma \not = 0\) and \(\sigma = 0\) respectively, where the constants \(k_1,\ldots, k_4\) are entirely determined numbers; if \(\sigma = 0\), \ \(\Omega_0 = \{|y| \le 1\}, \ \rho(0,y) = \rho_0(|y|), \ v(0,y) = v_0(|y|)\frac{y}{|y|},\) then the problem (1), (2) admits a global spherically symmetric solution, which is obtained in the explicit form, and domain \(\Omega_t\) is a ball with \(\mathrm{diam}\,\Omega_t = 2(1+v_0(1)t)\).