Subsonic potential flows in general smooth bounded domains (Q889112)

The paper under review deals with the existence and uniqueness of solutions of a certain class of PDEs. From the physical point of view, the solutions represent subsonic potential flows in a general bounded domain \(\Omega\) with smooth boundary (\(C^{2,\alpha}\)). From the mathematical point of view, the author studies a degenerate elliptic equation with Neumann boundary condition, actually derived from the Euler equations of steady isentropic compressible ideal flows. Assuming the usual Darcy-like law (i.e. the pressure is equal to a power \(\gamma>1\) of the density \(\rho\)) and that the velocity field is the gradient of a potential function, the main equation (for the potential function \(\varphi\) as unknown) reads as follows \[ \operatorname{div}(\rho(|\nabla\varphi|^2)\nabla\varphi)=0, \;\;in\;\Omega, \] equipped with the Neumann boundary condition \(\rho\nabla\varphi\cdot n=f\) on \(\partial\Omega,\) where \(n\) is the outward normal to \(\partial\Omega\) and \(f\in C^{1,\alpha}(\partial\Omega)\) is given, satisfying the compatibility condition \(\int_{\partial\Omega}f dS=0.\) Here \(\rho(|\nabla\varphi|^2)=\left(\frac{\gamma-1}{\gamma}(\bar B-\frac12 |\nabla\varphi|^2)\right)^{\frac{1}{\gamma-1}}\) is the density function depending on the potential flow and the Bernoulli constant~\(\bar B.\) The strategy to show the existence of a potential function satisfying the above problem relies on some classical truncation method, i.e. study first a non-degenerate equation for which the existence of a solution is well-known. Then proving some classical a priori estimates (for the truncated problem) the author shows the existence of a \(C^{2,\alpha}\) strong solution of the original problem (for any Bernoulli constant, larger than a threshold \(B^*\)). Uniqueness is obtained by the means of classical arguments. As a byproduct, the author shows that as \(\bar B\to B^*,\) the subsonic potential flow \(\nabla\varphi\) approaches to a subsonic-sonic flow that solves its corresponding Euler system in the sense of distributions (this is contained also in the main Theorem 1.1). The strategy here relies on a compensated compactness argument (in fact on the div-curl lemma of Murat and Tartar). In the reviewer's opinion, the presentation of the paper could have been improved, sometimes some details are missing and it is easy to lose focus while reading the paper.