Michigan State University
Gradient estimates for solutions to divergence form elliptic equations with piecewise constant coefficients in dimension N
Abstract
dc:descriptionIn a bounded C1,α0 domain \Omega \subset \mathbb{R}n, 0<α0\leq1, contains two simply connected and strictly convex C1,α0 subdomains (inclusions) D1 and D2 that satisfy D1\cup D2 \subset \subset \Omega and \overline{D1}\cap \overline{D2}=\lbrace 0 \rbrace, we study the following elliptic differential equation \begin{equation}\label{eq:eq2-intro-abs} \begin{split} \begin{cases} \text{div} \left(a(x)\nabla u\right)=0 \quad\text{in} \;\; \Omega,\\ \partial_{\nu}u(x)=g \qquad \qquad\text{on} \;\; \partial \Omega,\\ \int_{\partial \Omega} u=0, \end{cases} \end{split} \end{equation} where $a(x)=1+(k-1)\chi(D1\cup D2)(x),\quad 0<k<\infty, k\neq 1,$ and $g\in L20(\partial \Omega):=\lbrace g\in L2(\partial \Omega):\int\partial \Omega g=0\rbrace.$The problem arises from studying fiber re-inforced composite media with closely spaced inclusions in dimension $n=2$. For dimension $n=3$, the problem also appear in the study of steady state solutions to Maxwell's equations. The bound on the gradient yields the boundedness of the strain in fiber re-inforced materials and electrical fields when the two inclusions are touching each other. Following the ideas of \cite{Ammari:2015}, we assume that $D_1$ lies in the half space $x_{n}<0$, and $D_2$ in the half space $x_{n}>0$. Then we begin by separating the inclusions by a distance $\delta >0, that is, we set \Ddelta{1}=D1-\frac{δ}{2}\bold{en} , \text{and} \Ddelta{2}=D2+\frac{δ}{2}\bold{en}, $ where $\bold{e_n}=(0',1). Then we study the approximate differential equation corresponding to the separated inclusions which is \begin{equation}\label{eq:eq3-introduction-abs} \begin{split} \begin{cases} \text{div} \left(aδ(x)\nabla uδ\right)=0 \quad\text{in} \Omega,\ \partial\nuuδ(x)=g \qquad \qquad\text{on} \partial \Omega,\ \int\partial \Omega uδ=0,\ \end{cases} \end{split} \end{equation} where g\in L^2_0(\partial \Omega)$ and $aδ(x)=1+(k-1)\chi(D1δ\cup Dδ2)(x).$The solution of the elliptic equation (\ref{eq:eq3-introduction-abs}) has an integral representation in terms of potential functions defined on theboundary of each subdomain. From the representation formula, we derive uniform piecewise $C^{1,\alpha}$, $0<\alpha<\alpha_0, estimates for this solution which are independent of the distance between the subdomains. That is, we find the estimate \begin{align*} \norm{\ud}C1,α\Big(\Omega\tilde{ε}\setminus\overline{D1δ\cup D2δ }\Big)+ \norm{\ud}C1,α(\overline{D1δ })+\norm{\ud}C1,α(\overline{D2δ })\leq C \norm{g}L2(\partial \Omega), \end{align*} where \Omega_{\tilde{\epsilon}}=\lbrace x\in \Omega : \text{dist}(x,\partial \Omega)>\tilde{\epsilon} \rbrace$ and $C$ is independent of $\delta$. Our result extends the earlier result for dimension $n=2$ \cite{Ammari:2015}, but the analysis is much more complicated. Final estimates rely on detailed analysis near the touching point and collective compactness of some integral operators.
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Al-Yasiri, Khaldoun
- Contributors dc:contributor
-
- Zhou, Zhengfang
- Abbas, Casim
- Yan, Baisheng
- Tang, Moxun
- Wong, Willie
Subjects
dc:subject × 3Rights
dc:rights- Statement dc:rights
-
- In Copyright
- Language dc:language
- English
Identifiers
dc:identifier.*- Identifier
-
etd:4411
isbn:9781339949833
isbn:1339949830
oclc:961332699
umi:10139497
local:AlYasiri_grad.msu_0128D_14676 - OAI record oai:identifier
- https://d.lib.msu.edu/oai?verb=GetRecord&metadataPrefix=oai_dc&identifier=oai%3Ad.lib.msu.edu%3Aetd_4411