On periodic homogenization in perfect elasto-plasticity (Q2441486): Difference between revisions

The authors describe a homogenization result for a material composed of a finite number of elasto-plastic phases by extending their previous result [Commun. Pure Appl. Math. 65, No. 9, 1185--1241 (2012; Zbl 1396.74036)]. They consider an open bounded domain \(\Omega \) in \(\mathbb{R}^{N}\) with Lipschitz boundary \(\partial \Omega \). Dirichlet boundary conditions are imposed on a part \(\Gamma _{D}\) of \(\partial \Omega \), and an admissibility condition is imposed on the boundary of \(\Gamma _{D}\) in \(\partial \Omega \). Let \(\mathcal{Y} =\mathbb{R}^{N}/\mathbb{Z}^{N}\) be the \(N\)-dimensional torus. \(\Omega \) is composed of a finite number of phases \((\mathcal{Y}_{i})_{\varepsilon }=\{x\in \mathbb{R}^{N}:x/\varepsilon \in \mathcal{I}(\mathcal{Y}_{i})+ \mathbb{Z}^{N}\}\), where \(\mathcal{I}:\mathcal{Y}\rightarrow [ 0,1)^{N} \) is the canonical identification associated to the torus \(\mathcal{Y}\). The elasticity tensor associated to these phases is given as \(\mathbb{ C}(y)M=\mathbb{C}_{D}(y)M_{D}+k(y)\mathrm{tr}(M)\mathfrak{i}\), for every \(y\in \mathcal{Y}\), where \(\mathbb{C}_{D}\in \mathcal{L}_{s}(M_{D}^{N})\), \( M_{D}^{N}\) being the set of free-trace and symmetric matrices, \(k(y):=k_{i}>0 \) on each phase \(\mathcal{Y}_{i}\). Then the authors define the admissible conditions introducing a convex and compact set \(K_{i}\subset M_{D}^{N}\) for each phase \(\mathcal{Y}_{i}\) and imposing boundedness conditions on \(K_{i}\). They also define the dissipation potential \(H(y,\xi )=H_{i}(\xi ):=\sup \{\tau \cdot \xi :\tau \in K_{i}\}\) in the phase \( \mathcal{Y}_{i}\). The problem under consideration is called a quasi-static evolution involving the displacement field \(u_{\varepsilon }\), an elastic strain \(e_{\varepsilon }\) and a plastic strain \(p_{\varepsilon }\), which must satisfy a global stability condition and an energy inequality. The authors first prove an existence result for a quasi-static evolution starting from an initial data which satisfy the global stability condition. For the description of their homogenization result, the authors essentially use the two-scale convergence, but they here extend the original definition of this tool (first defined by \textit{G. Nguetseng} [SIAM J. Math. Anal. 20, No. 3, 608--623 (1989; Zbl 0688.35007)]) to the case of measures. They prove the main properties of this extended notion of two-scale convergence, among which is a compactness result in the set \(\mathcal{M}_{b}(\Omega \times \mathcal{Y})\) of Radon measures in \(\Omega \) with values in \(\mathcal{ Y}\). The authors also use an unfolding result as introduced by \textit{D. Cioranescu} et al. [SIAM J. Math. Anal. 40, No. 4, 1585--1620 (2008; Zbl 1167.49013)]. Then they characterize the set of admissible two-scale configurations. The paper ends with the description of properties of two-scale quasi-static evolutions. These indeed satisfy a flow rule property.