A first order in time wave equation modeling nonlinear acoustics (Q7031709)

The authors consider the Navier-Stokes-Fourier system that describes the motion of a viscous and heat conducting fluid: \(\partial_{t}\rho +\nabla \cdot (\rho v)=h\), \(\rho \lbrack \partial_{t}v+(v\cdot \nabla)v]=-\nabla p+\mu_{S}\Delta v+(\frac{\mu_{S}}{3}+\mu_{B})\nabla (\nabla \cdot v)+g\), \(\rho \theta \lbrack \partial_{t}s+(v\cdot \nabla)s]=\kappa \Delta \theta + \frac{\mu_{S}}{2}(\sum_{i,j,k}\partial_{x_{j}}v_{i}+\partial_{x_{i}}v_{j}- \frac{2}{3}\delta_{ij}\partial_{x_{k}}v_{k})+\mu_{B}(\nabla \cdot v)^{2}+\ell\), where \(\rho\) is the mass density, \(v\) the velocity field, \(S\) the entropy, \(p\) the pressure, \(\theta\) the temperature of the acoustic medium, \(\delta_{ij}\) the Kronecker-Delta. symbol, \(\mu_{S}\) the shear viscosity, \(\mu_{B}\) the bulk viscosity, and \(\kappa\) the thermal conductivity.\ The initial conditions \(v_{0}=0\), \(\partial_{t}\rho_{0}=\partial_{t}s_{0}=\partial_{t}p_{0}=\partial_{t}\theta_{0}=0\), \(\nabla \rho_{0}=\nabla s_{0}=\nabla p_{0}=\nabla \theta_{0}=0\) are imposed. \N\NThe authors consider sound waves with small amplitude that travel through some medium and they suppose that the Mach number \(\varepsilon\) and acoustic Reynolds number \(\eta\) are small.\ They perform a change of variables in the preceding system and introduce a linear perturbation in terms of the Mach number in every unknown, leading to dimensionless equation.\ They finally introduce a linear approximation ending with the system: \(\partial_{t^{\prime}}p^{\prime}+\nabla^{\prime}\cdot v^{\prime}=0\), \(\partial_{t^{\prime}}v^{\prime}+\nabla^{\prime}p^{\prime}=0\), \(\partial_{t^{\prime}}s^{\prime}=0\), \(\rho^{\prime}=p^{\prime}-s^{\prime}\), \(\theta^{\prime}=p^{\prime}+\frac{s^{\prime}}{\gamma -1}\), up to \(O(\varepsilon, \eta)\) terms, where \(\gamma\) denotes the heat capacity ratio. They also introduce a quadratic approximation and end with the system: \(\partial_{t}p-\mu \Delta p+(1+\gamma)\nabla \cdot v+\varepsilon_{1}p\nabla \cdot v+\varepsilon_{2}\nabla p\cdot v=h\), \(\partial_{t}v-\eta \Delta v+(1+\delta)\nabla p-\frac{\varepsilon^{3}}{2} \nabla p^{2}+\frac{\varepsilon^{4}}{2}\nabla v^{2}=g\), posed in \(\Omega \times (0,T)\), where \(\Omega\) is a bounded \(C^{1,1}\)-regular domain in \(\mathbb{R}^{d}\), \(d=2,3\) and \(T\) is finite. The initial conditions \(p(0)=p_{0}\), \(v(0)=v_{0}\) and the boundary conditions \(p=0\), \(v=0\) are imposed on \(\partial \Omega\). \N\NThe authors prove the following existence and uniqueness result: there exist positive \(r,\widetilde{r}\) (small enough) and positive \(C_{G}\) independent of \(T\) such that for each \(u_{\ast}\in B_{r}^{X}(0)\), \((f,u_{0})\in Y=L^{2}(T;(L^{2})^{n})\times (H_{0}^{1})^{n}\) and \(\gamma, \delta \in B_{r}^{L^{\infty}}\), there exists a unique \(u\in X=H^{1}(T;(L^{2})^{n})\cap L^{2}(T;(H^{2}\cap H_{0}^{1})^{n})\cap L^{\infty}(T;(H^{1})^{n})\), with \(\mathcal{F}_{u_{\ast}}^{\prime}[u]=(f,u_{0})\), \(\mathcal{F}_{u_{\ast}}^{\prime}\) being the Fréchet derivative at \(u_{\ast}\in X\) of the operator \(\mathcal{F}:X\rightarrow Y\) associated with the above equations and initial conditions.\ Moreover \(u\) satisfies the a priori estimate \(\left\Vert u\right\Vert_{X}\leq C_{G}\left\Vert (f,u_{0})\right\Vert_{Y}\). For the proof, the authors use the Galerkin approximation method and they prove uniform estimates on the sequence of approximate solutions, using Newton-Kantorovich theorem. The second main result proves the existence of a global-in-time solution and an exponential decay of this solution, under further hypotheses on the data. They finally consider the limit problem \(\partial_{t}p-\mu \Delta p+\nabla \cdot v=f\), \(\partial_{t}v-\eta \Delta v+\nabla p=g\), that they write in an abstract way as: \(u^{\prime}(t)=\mathbb{A}_{0}u(t)\). They decompose the operator \(\mathbb{A}_{0}\) into a skew adjoint and a self-adjoint positive definite part. They prove that \(\mathbb{A}_{0}\) is the infinitesimal generator of a contraction semigroup which is analytic and exponentially stable.