Optimal control of parameters for elastic body with thin inclusions (Q2302743): Difference between revisions

The authors consider a bounded domain \(\Omega \subset \mathbb{R}^{2}\) with Lipschitz boundary \(\Gamma \), a smooth curve \(a_{r}\) such that \(\overline{ a_{r}}\subset \Omega \) and \(a_{e}=(-1,0)\times \{0\}\), assuming that \( a=a_{e}\cup a_{r}\cup \{0,0)\}\) is a smooth curve with a 0 angle between \( a_{e}\) and \(a_{r}\) at \((0,0)\). \(a_{e}\) and \(a_{r}\), respectively, correspond to thin elastic and rigid inclusions incorporated in the elastic material which fills in \(\Omega _{a}=\Omega \setminus \overline{a}\). The authors consider the Euler-Bernoulli beam model for \(a_{e}\) and assume that the delamination takes place at the positive side of \(a_{e}\). They define on \( H^{1}(\Omega _{a_{e}})^{2}\times H^{2}(a_{e})\times H^{1}(a_{e})\) the energy functional \[E_{\delta }(u,v,w)=\frac{1}{2}\int_{\Omega _{a}}\sigma (u)\cdot \varepsilon (u)-\int_{\Omega _{a}}g\cdot u+\frac{1}{2} \int_{a_{e}}v_{,11}^{2}+\frac{1}{2}\int_{a_{e}}w_{,1}^{2}+\frac{1}{2\delta } (\rho _{2,1}(0+)-v_{,1}(0-))^{2},\] where \(g=(g_{1},g_{2})\in L^{2}(\Omega )^{2}\) are given external forces, \(\rho =(\rho _{1},\rho _{2})=u\mid _{a_{r}} \), \(\varepsilon (u)=\{\varepsilon _{ij}(u)=\frac{1}{2}(u_{i,j}+u_{j,i})\}\) is the strain tensor and \(\sigma (u)=C\varepsilon (u)\) is Hooke's law, \(C\) being the elasticity tensor which satisfies the usual symmetry and positive definite properties. They introduce the set of admissible displacements as \( P=\{(u,v,w)\in H^{1}(\Omega _{a_{e}})^{2}\times H^{2}(a_{e})\times H^{1}(a_{e}):[u_{\nu }]\geq 0\),\(\ v=u_{\nu }^{-}\), \(w=u_{\tau }^{-}\) on \( a_{e}\), \(u\mid _{a_{r}}\in R(a_{r})\}\), with \(R(a_{r})=\{\rho :\rho (x_{1},x_{2})=b(-x_{2},x_{1})+(c^{1},c^{2})\), \((x_{1},x_{2})\in a_{r}\}\) for arbitrary constants \(b\), \(c^{1}\), \(c^{2}\). To the minimization problem \( \inf_{P}E_{\delta }\), they associate a differential formulation. They prove the equivalence between these formulations, assuming that the solution is smooth and using the variational formulation: \(\int_{\Omega _{a}}\sigma (u)\cdot \varepsilon (\overline{u}-u)-\int_{\Omega _{a}}g\cdot (\overline{u} -u)+\int_{a_{e}}v_{,11}(\overline{v}_{,11}-v_{,11})+\int_{a_{e}}w_{,1}( \overline{w}_{,1}-w_{,1})+\frac{1}{\delta }(\rho _{2,1}(0+)-v_{,1}(0-))( \overline{\rho }_{2,1}(0+)-\overline{v}_{,1}(0-)-\rho _{2,1}(0+)+v_{,1}(0-))\geq 0\), for every \((\overline{u},\overline{v}, \overline{w})\in P\). Letting \(\delta \) increase to \(+\infty \), the authors prove the existence of \((u,v,w)\) such that \(u^{\delta }\rightharpoonup u\) weakly in \(H^{1}(\Omega _{a_{e}})^{2}\), \(u\mid _{a_{r}}=\rho ^{0}\in R(a_{r}) \), \((v^{\delta },w^{\delta })\rightharpoonup (v,w)\) weakly in \( H^{2}(a_{e})\times H^{1}(a_{e})\) and \((u,v,w)\in P\) is the solution of the variational inequality \(\int_{\Omega _{a}}\sigma (u)\cdot \varepsilon ( \overline{u}-u)-\int_{\Omega _{a}}g\cdot (\overline{u}-u)+ \int_{a_{e}}v_{,11}(\overline{v}_{,11}-v_{,11})+\int_{a_{e}}w_{,1}(\overline{ w}_{,1}-w_{,1})\geq 0\), for every \((\overline{u},\overline{v},\overline{w} )\in P\). The authors then consider the case where \(\delta \) goes to 0 and they describe the limit problem. In a further part of their paper, the authors add a parameter \(\lambda \) associated to the rigidity of the elastic inclusion and they describe the asymptotic behavior of the solution when \( \lambda \rightarrow \infty \). In the last part of their paper, the authors consider an optimal control problem associated to this context for which they prove an existence result. They also prove that the inverse problem has a solution.