Optimal density lower bound on nonisentropic gas dynamics (Q2286701)

The author of this paper investigates a time-dependent lower bound on the density of optimal order \(O(1/(1 + t))\) for the general smooth nonisentropic flow of compressible Euler equations represented in Lagrangian coordinates in one space dimension, \[ (a) \ \ \tau_t-u_x=0, \ \ (b) \ \ u_t+p_x=0, \ \ (c) \ \ (u^2/2+e)_t+(up)_x=0, \] where \(\rho \) is the density, \(\tau = \rho^{-1}\) is the specific volume, \(p\) the pressure, \(u\) the velocity, \(e\) the specific internal energy, \(t\in\mathbb{R}^{+}\) is the time and \(x\in\mathbb{R}\) is the spatial coordinate. The system is closed by the second law of thermodynamics. Note that for any \(C^1\)-solution, it follows that the equation \((c)\) is equivalent to the conservation of entropy \(S_t=0\) (\(S(x,t)\equiv S(x,0) =: S(x)\)). If the entropy is constant, the flow is isentropic, then the pair of equations \((a)\) and \((b)\) become a closed system, known as the \(p\)-system. Next the polytropic ideal gas is considered, that is, \(p\tau =RT\), \(e=c_vT=p\tau /(\gamma - 1)\), where \(R>0\) (ideal gas constant), \(c_v>0\) (specific heat constant). It implies that the additional equality, \ \ \((d)\) \ \(p=Ke^{S/c_v}\tau^{-\gamma }\) hold with adiabatic gas constant \(\gamma > 1\) (\(K>0\)). The main result in short is that for any classical nonisentropic solution, the density has a lower bound of order \(O(1 + t)^{-1} \). Actually the main statement is formulated in the following way. The author introduces the functions \(\tilde{\alpha }\) and \(\tilde{\beta }\) depending explicitly on \(u\) and pair \(m\) and \(\eta \) (\(\eta_x\)) which are some new variables instead of \(S\) and \(\tau \), respectively. Consider a \(C^1\) solution \((u(x,t),\tau (x,t),S(x))\) of the initial value problem for the compressible Euler equations \((a), \ (b)\), \(S_t=0\) and the equality \((d)\) \ \ (\((x,t)\in \mathbb{R}\times [0,T)\), \(T\) can be any finite positive constant or \(+\infty \)), under the initial condition \((u(x, 0),\tau (x, 0),S(x)) \in C^1\cap L^{\infty }(\mathbb{R})\). Also assume that the initial data \(\rho (x,0)=1/\tau (x,0)\), \(S^{\prime }(x)\), \(\tilde{\alpha }(x,0)\), \(\tilde{\beta }(x,0)\) are all uniformly bounded for any \(x\). The total variation of \(S(x)\) is finite. Then \(\min\limits_{x\in\mathbb{R}}\rho (x,t)>C_1(1+t)^{-1}\), where \(C_1\) is a special constant depending on the \(C^1\cap L^{\infty }(\mathbb{R})\) norm of the initial data \((u,\tau ,\rho ,S)(x, 0)\), but it is independent of \(T\). Moreover, \(\tilde{\alpha }(x,t)\) and \(\tilde{\beta }(x,t)\) are both bounded above by constants independent of \(T\), \((x, t)\in \mathbb{R}\times [0,T)\). The proof of this statement is built on the observation of a new invariant domain for some gradient variables, given in another statement which itself is also a result of interest for the reader.