Convergence rates of Neumann problems for Stokes systems

From MaRDI portal
Publication:2405372



Abstract: This paper studies the convergence rates in L2 and H1 of Neumann problems for Stokes systems with rapidly oscillating periodic coefficients, without any smoothness assumptions on the coefficients.


The author proves convergence rate results for a homogenization result starting with a Stokes system with a Neumann boundary condition. The author considers a bounded and \(C^{1,1}\) domain \(\Omega \subset \mathbb{R}^{d}\) and the Stokes system \(\mathcal{L} _{\varepsilon }(u_{\varepsilon })+\nabla p_{\varepsilon }=F\), \(\operatorname{div}(u_{\varepsilon })=g\) in \(\Omega \). Here \(\mathcal{L}_{\varepsilon }=-\operatorname{div}(A(x/\varepsilon )\nabla )\), where the matrix \(A\) is 1-periodic and satisfies uniform continuity and coercivity conditions. The Neumann boundary condition \(\frac{\partial u_{\varepsilon }}{\partial \nu _{\varepsilon }} -p_{\varepsilon }n_{\varepsilon }=f\) is imposed on \(\partial \Omega \). \(F\in L^{2}(\Omega ;\mathbb{R}^{d})\) and \(f\in H^{1/2}(\partial \Omega ;\mathbb{R} ^{d})\) satisfy the compatibility condition \(\int_{\Omega }F+\int_{\partial \Omega }f=0\) and \(g\in H^{1}(\Omega )\). The first main result proves that \( \left\| u_{\varepsilon }-u_{0}\right\| _{L^{2}(\Omega )}\leq C\varepsilon \left\| u_{0}\right\| _{H^{2}(\Omega )}\), where \( u_{\varepsilon }\) (resp. \(u_{0}\)) is the unique weak solution to the original (resp. homogenized) Stokes problem. The second main result proves that \(\left\| u_{\varepsilon }-u_{0}-\varepsilon \chi ^{\varepsilon }S_{\varepsilon }(\nabla \widetilde{u_{0}})\right\| _{H^{1}(\Omega )}\leq C\sqrt{\varepsilon }\left\| u_{0}\right\| _{H^{2}(\Omega )}\), where \(\chi ^{\varepsilon }\) is associated to the corrector \((\chi ,\pi )\), \(S_{\varepsilon }\) is the Steklov smoothing operator and \(\widetilde{u_{0}}\) is the extension to \(\mathbb{R}^{d}\) of \( u_{0}\). The author here assumes that \(\int_{\Omega }u_{\varepsilon }=\int_{\Omega }u_{0}=0\). A similar convergence rate result is proved for the pressure in \(L^{2}(\Omega )\). The author first proves the existence of a unique weak solution to the original Stokes problem, using a variational formulation. Then he recalls the homogenization result, he proved in his PhD dissertation [Homogenization of Stokes systems with periodic coefficients. Lexington, KY: Univ. Kentucky (PhD Thesis) (2016; \url{doi:10.13023/ETD.2016.302})] for this Stokes problem, using Tartar's oscillating testing function method. He also recalls the structure of the corrector, and he presents dual correctors. He introduces the Steklov smoothing operator defined on \(L^{2}(\mathbb{R}^{d};\mathbb{R}^{d})\) through \((S_{\varepsilon }u)(x)=\int_{Y}u(x-\varepsilon z)dz\), and he proves that \(\left\| S_{\varepsilon }u-u\right\| _{L^{2}(\mathbb{R}^{d})}\leq C\left\| \nabla u\right\| _{L^{2}(\mathbb{R}^{d})}\) for every \(u\in H^{1}(\mathbb{R }^{d})\). For the proof of the convergence rate results, the author uses the variational formulation with appropriate test functions and he quotes Suslina's estimates for integrals close to the boundary \(\partial \Omega \).



Cites work









This page was built for publication: Convergence rates of Neumann problems for Stokes systems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2405372)