A variational approach to the local character of \(G\)-closure: the convex case (Q1012351)

The problem of the determination of the \(G\)-closure is here raised within the nonlinear elasticity framework. The authors consider \(N\) energy densities \( W^{(i)}\), \(i=1,\dots ,N\), which are continuous functions belonging to the set \(\mathcal{F}(\alpha ,p)\), that is satisfying a \(p\)-growth and a \(p\)-coercivity condition, with \(p>1\). A composite material is obtained as a mixture of these different basic materials considering the energy density \( W_{\theta }=\sum_{i}\theta _{i}W^{(i)}\), for some functions \(\theta _{i}\in L^{\infty }(\Omega ;[0,1])\), satisfying \(\sum_{i}\theta _{i}=1\), a.e. in \( \Omega \). One main result of the paper characterizes the \(G\)-closure. Every effective energy density in this \(G\)-closure can be locally recovered through a \(\Gamma \)-convergence process as the pointwise limit of the sequence of periodic homogenized energy densities with prescribed volume functions. This is more precisely obtained taking a sequence \((\chi _{k})\) which satisfies \(\sum_{i}\chi _{k}^{(i)}=1\), a.e. in \(\Omega \), where \(\chi _{k}(y)=\chi (\langle ky\rangle )\) and which converges to \( \theta \) in the weak\(^{\ast }\) topology of \(L^{\infty }(\Omega ;[0,1]^{N})\). The authors first claim that they can assume the quasiconvexity of the energy densities. The proof of this result is based on the properties of the \(\Gamma \)-convergence (for the weak topology of \(W^{1,p}\)) and on classical functional analysis tools. When removing the convexity condition, but assuming that the zero-level set of the effective energy density can be associated to Young measures, the authors prove some similar characterization of the \(G\)-closure. The authors also discuss the behaviour of cell integrands. The paper ends with some counter-examples to the results or properties exposed in the results obtained throughout the paper. One of these counter-examples deals with the possibility to obtain similar results when considering a two-scale convergence method.