Mathematical theory of evolutionary fluid-flow structure interactions (Q1650000)

The book gathers four lectures held in Oberwolfach in 2016 and dealing with the mathematical analysis of different kinds of fluid-structure interactions. In Chapter 1, I. Kukavica and A. Tuffaha describe a general fluid-structure interaction, assuming that the structure moves within the fluid with a frictional damping. They start writing the fluid motion in 3D Eulerian coordinates through Navier-Stokes equations in a domain \(\Omega _{f}(t)\) which depends on time starting from the flat domain \(\Omega _{f}=\{x\in \mathbb{R}^{3}:0\leq x_{3}\leq 1\) or \(2\leq x_{3}\leq 3\}\). The linear elastic material fills the domain \(\Omega _{e}=\{x\in \mathbb{R} ^{3}:1\leq x_{3}\leq 2\}\) and its motion is described by the hyperbolic equation \(\frac{\partial ^{2}w}{\partial t^{2}}-div_{y}\sigma (w)=0\) with \(\sigma (w)=\lambda Tr\epsilon (w)I+2\mu \epsilon (w)\). The authors then move to a Lagrangian formulation of the whole problem as \( v_{t}^{k}-\partial _{j}(a_{l}^{j}a_{l}^{i}\partial _{i}v^{k})+\partial _{i}(a_{k}^{i}q)=0\), \(a_{k}^{i}\partial _{i}v^{k}=0\) in \(\Omega _{f}\times (0,T)\), \(w_{tt}^{k}-\Delta w^{k}+\alpha w_{t}^{k}+\beta w^{k}=0\) in \(\Omega _{e}\times (0,T)\). The continuity of the velocities \(w_{t}^{k}=v^{k}\) and of the stresses \(\partial _{j}w^{k}N_{j}=a_{l}^{j}a_{l}^{i}\partial _{t}v^{k}N_{j}-a_{k}^{i}qN_{i}\) is imposed on the common boundary \(\Gamma _{c}\times (0,T)\).\ On the rest of the boundary, the fluid is supposed to be at rest. Initial conditions are imposed to \(v\), \(w\) and \(w_{t}\). The authors first quote from the paper by \textit{P. B. Mucha} and \textit{W. Zajaczkowski} in [Stud. Math. 143, No. 1, 75--101 (2000; Zbl 0970.35107)] the existence of a global in time weak solution to a nonhomogeneous Stokes problem with given divergence. They assume that the initial condition \(v_{0}\in H^{1}(\Omega _{f})\) satisfies compatibility conditions. The authors prove regularity results for this Stokes problem. They then consider a nonhomogeneous Stokes problem with variable coefficient \(v_{t}^{k}-\partial _{j}(a_{l}^{j}a_{l}^{i}\partial _{i}v^{k})+\partial _{i}(a_{k}^{i}q)=f^{k}\), \(a_{k}^{i}\partial _{i}v^{k}=g\) in \(\Omega _{f}\times (0,T)\), with the mixed boundary conditions \(a_{l}^{j}a_{l}^{i}\partial _{t}v^{k}N_{j}-a_{k}^{i}qN_{i}=h^{k}\) on \(\Gamma _{c}\times (0,T)\) and \(v=0\) on \(\Gamma _{f}\times (0,T)\). The existence of a unique weak solution to this problem is ensured if the initial condition \(v_{0}\in H^{1}(\Omega _{f}) \) satisfies compatibility conditions and if the coefficients \(a_{k}^{i}\) satisfy the Piola condition \(\partial _{i}a_{k}^{i}=0\) and smallness properties for \(I-a\) and \(\partial _{t}a\). They then move to the static fluid-solid system with constant coefficients \(v_{t}-\Delta v+\nabla q=f\), \( div\;v=g\) in \(\Omega _{f}\times (0,T)\), \(w_{tt}-\Delta w=-\alpha w_{t}-\beta w=0\) in \(\Omega _{e}\times (0,T)\), with the boundary conditions \( v=w_{t}\), \(\frac{\partial v}{\partial N}-qN=\frac{\partial w}{\partial N}+h\) on \(\Gamma _{c}\times (0,T)\), \(v=0\)\ on \(\Gamma _{f}\times (0,T)\). Initial conditions are imposed on \(v\), \(w\) and \(w_{t}\). The authors assume regularity properties for the initial conditions and the data. They prove the existence of a unique weak solution. The authors finally prove the existence of a unique weak solution to the original fluid-structure system. The talk ends with an equipartition result for the energy associated to this system. The second lecture by R. Triggiani deals with a linear parabolic-hyperbolic fluid-structure model posed in a bounded and open domain \(\Omega \subset \mathbb{R}^{d}\), \(d=2,3\), which is decomposed as \(\Omega =\Omega _{f}\cup \Omega _{s}\cup \Gamma _{s}\) and where the fluid part \(\Omega _{f}\) surrounds the solid part \(\Omega _{s}\), and written in a simplified form as \( u_{t}-\Delta u+\nabla p=0\), \(div\;u=0\) in \((0,T]\times \Omega _{f}\), \( w_{tt}-\Delta w+w_{t}+\chi w=0\) in \((0,T]\times \Omega _{s}\). The boundary conditions \(u\mid _{\Gamma _{f}}=0\) on \((0,T]\times \Gamma _{f}\), \( u=w_{t}-\alpha \frac{\partial w}{\partial \nu }\), \(\frac{\partial u}{ \partial \nu }-\frac{\partial w}{\partial \nu }=p\nu \) on \((0,T]\times \Gamma _{f}\) are imposed. Initial conditions are imposed on \(u\), \(w\) and \( w_{t}\). The damping coefficients \(\alpha \) and \(\chi \) will only take the two values 0 and 1. The first part of the lecture deals with the undamped case \(\chi =\alpha =0\). The author writes the problem in an abstract linear way through linear maps \(G_{1}\) and \(G_{2}\) and an operator \(\mathcal{A}\). He quotes from papers mainly written with S. Avalos properties of the operator \(\mathcal{A}\), of its point spectrum \(\sigma _{p}(\mathcal{A})\) and of the adjoint \(\mathcal{A}^{\ast }\) and this first part ends with higher regularity results for the solution. The second part of the lecture deals with the interior damped case \(\chi =1\), \(\alpha =0\). The author here again writes the corresponding problem in an abstract linear way through through linear maps \(G_{1}\) and \(G_{2}\) and an operator \(\mathcal{A}\). He proves that the operator \(\mathcal{A}\) generates a semigroup on the set \(\mathcal{H} \) of initial data and a uniform stabilization property of \(\widehat{\mathcal{ A}}\) on \(\widehat{\mathcal{H}}\). The third part of the lecture deals with the case of dissipation on the interface, \(\chi =0\), \(\alpha =1\). The author here proves properties of the associated operator \(\mathcal{A}_{F}\): generation of a strongly stable contraction semigroup on \(\mathcal{H}\) and properties of the spectrum. The lecture ends with three appendices concerning regularity results for the nonhomogeneous Stokes problem, further computations on the systems under consideration and on the overdetermined elliptic problem \(\Delta \varphi =-\mu \varphi \) in \(\Omega \) with boundary conditions \(\varphi \mid _{\partial \Omega }=0\) and \(\frac{\partial \varphi }{\partial \nu }=k\) on an open connected subportion \(\Gamma _{0}\) of \( \partial \Omega \), where \(\mu >0\) and \(k\) is an unspecified constant. The author analyses which geometries satisfy or not this problem. The third lecture is given by I. Lasiecka and J.T. Webster and deals with flow-plate interactions, especially focusing on the role of the underlying semigroup in the long-time behavior of the solution. The lecture starts with a general presentation of dynamical systems and their long-time behavior. The authors then analyze the flutter phenomenon in the case of flow-plate interaction problem written as \((1-\alpha \Delta )u_{tt}+\Delta ^{2}u-D_{0}(\nabla u_{t},u_{t})+f(u)=p(x,t)\) posed in \(\Omega \times (0,T)\) . Boundary conditions are imposed on \(u\) on the boundary \(\partial \Omega \) and the solution starts from initial data \(u(0)=u_{0}\) and \(u_{t}(0)=u_{1}\). The authors first consider the case of a von Karman nonlinearity \( f_{V}(u)=-[u,v(u)+F_{0}]\) where \(F_{0}\) is the in-plane load and \(v\) is the Airy stress function solution of \(\Delta ^{2}v(u_{1},u_{2})+[u_{1},u_{2}]=0\) in \(\Omega \) with the boundary conditions \(\partial _{\nu }v(u_{1},u_{2})=0=v(u_{1},u_{2})\) on \(\partial \Omega \). The authors quote from the book by \textit{I. Chueshov} and \textit{I. Lasiecka} [Springer Monographs in Mathematics. New York, NY: Springer (2010; Zbl 1298.35001)] a local Lipschitz character of the von Karman nonlinear function. They then move to a panel flow-plate interaction problem written as \(M_{\alpha }u_{tt}+\Delta ^{2}u+k_{0}M_{\alpha }u_{t}+f_{V}(u)=p_{0}+[\partial _{t}+U\partial _{x}]\phi \mid _{\Omega }\) posed in \(\Omega \times (0,T)\) with the initial conditions \(u(0)=u_{0}\) and \( u_{t}(0)=u_{1}\) and the boundary conditions \(u=\partial _{\nu }u=0\) on \( \partial \Omega \times (0,T)\), \((\partial _{t}+U\partial _{x})^{2}\phi =\Delta \phi \) in \(\mathbb{R}_{+}^{3}\times (0,T)\) with the initial conditions \(\phi (0)=\phi _{0}\) and \(\phi _{t}(0)=\phi _{1}\) and the boundary conditions \(\partial _{z}\phi =-[(\partial _{t}+U\partial _{x})u]_{ext}\) on \(\{z=0\}\times (0,T)\). The authors here precise the functional setup adapted to this problem and they define the notion of solutions they will consider. Assuming hypotheses on the data, they prove the existence of a locally Lipschitz semigroup solution in the subsonic case. In the supersonic case, they slightly modify the problem and they again prove the existence of a semigroup solution. Then the authors describe the long-time behavior of flow-plate interactions. The lecture ends with some further considerations on such problems and with the presentation of open problems. The final lecture is given by B. Kaltenbacher and deals with nonlinear acoustics in the case of structure-acoustic coupling. The author starts with the nonlinear acoustic problem written as \(u_{tt}-c^{2}\Delta u-b\Delta u_{t}=\frac{k}{2}(u^{2})_{tt}=kuu_{tt}+k(u_{t})^{2}\) in the Westervelt formulation, posed in \(\Omega \times (0,T)\) where \(\Omega \) is a bounded domain in \(\mathbb{R}^{d}\), \(d=1,2,3\), and the associated energy \(E[u](t)= \frac{1}{2}(\left| u_{tt}(t)\right| _{L^{2}(\Omega )}^{2}+\left| \nabla u_{t}(t)\right| _{L^{2}(\Omega )}^{2}+\left| \Delta u(t)\right| _{L^{2}(\Omega )}^{2})\). The author quotes from the paper she wrote with \textit{I. Lasiecka} in [Discrete Contin. Dyn. Syst., Ser. S 2, No. 3, 503--523 (2009; Zbl 1180.35108)] the existence of a unique solution to this problem, its energy remaining bounded and the solution being nondegenerate and with exponential decay, assuming that the initial energy \( E[u](0)\) is less than some \(\tau >0\). The author then analyzes acoustic-acoustic or acoustic-elastic couplings which have applications in different contexts. The domain \(\Omega \) is decomposed in \(\Omega _{+}\cup \Omega _{-}\cup \Gamma \), \(\Omega _{-}\) being the acoustic part, \(\Omega _{+} \) the acoustic or elastic part and \(\Gamma \) the interface. The author introduces the following modified and damped equation \((1-ku)u_{tt}-c^{2} \Delta u-b\nabla \cdot ((1+\delta \left| \nabla u_{t}\right| ^{p-1}\nabla u_{t})=k(u_{t})^{2}\), with homogeneous Dirichlet boundary conditions and inhomogeneous initial conditions \(u(0)=u_{0}\), \(u_{t}(0)=u_{1} \). She quotes from the paper written by \textit{R. Brunnhuber}, \textit{B. Kaltenbacher} and \textit{P. Radu} in [Evol. Equ. Control Theory 3, No. 4, 595--626 (2014; Zbl 1304.35434)] the existence of a weak solution to this problem assuming that \(p>d-1\), \(p\geq 1\) and that the initial value \( u_{0},u_{1}\in W_{0}^{1,p-1}(\Omega )\) are small enough. This solution is unique in some ball of the space induced by the initial data. She then derives the existence of a solution to an acoustic-acoustic problem and to an acoustic-elastic problem. The last part of this lecture is devoted to a shape optimization problem in the context of lithotripsy, for which the author proves an existence result. These lectures will give to the reader involved in such problems the current available results and the tools which can be applied to treat them.