From finite to linear elastic fracture mechanics by scaling (Q2510349)

The authors describe the asymptotic behaviour of the propagation of a crack in a 2D brittle material. Let \(\Omega \) be a bounded, open and connected domain in \(\mathbb{R}^{2}\) whose Lipschitz boundary \(\partial \Omega \) may be decomposed as the disjoint union \(\partial _{D}(\Omega )\cup \partial _{N}(\Omega )\), with positive Hausdorff measure \(\mathcal{H}^{1}(\partial _{D}\Omega )\). The authors assume that \(K_{L}=(0,L]\times \{0\}\subset \Omega \), \(\partial _{D}(\Omega )\subset \partial \Omega \backslash ([0,L]\times \{0\})\), that the initial crack is given as \( K_{0}=(0,l_{0}]\times \{0\}\subset \Omega \) and that the admissible cracks are \(K_{\ell }=(0,\ell ]\times \{0\}\) for \(\ell \in [ l_{0},L]\). They consider a hyperelastic energy functional \(W:\mathbb{R}^{2\times 2}\rightarrow [ 0,+\infty ]\) which is polyconvex and \(C^{2}\) in the set \(\mathbb{R}_{+}^{2\times 2}\) of \(2\times 2\) matrices with positive determinants. In the first part of their paper, the authors describe some properties of such energy functionals and two examples: Ogden and Mooney-Rivlin materials. They then describe the quasi-static evolutions introducing the set of admissible displacements as \(\mathcal{U}(t,\ell )=\{u\in H^{1}(\Omega \backslash K_{\ell };\mathbb{R}^{2}):u=\alpha (t)g\) on \(\partial _{D}(\Omega )\}\) for \(\alpha \in C^{1,1}([0,T])\) and \(g\in W^{1,\infty }(\Omega ;\mathbb{R}^{2})\) and the nonlinear energy \(\mathcal{F}\) on \([0,T]\times [ l_{0},L]\times H^{1}(\Omega \backslash K_{\ell }; \mathbb{R}^{2})\) through \[ \mathcal{F}(t,\ell ,u)=\begin{cases} \int_{\Omega \backslash K_{\ell }}W(I+Du)dx & \text{if }u\in \mathcal{U} (t,\ell ), \\ +\infty & \text{otherwise.} \end{cases} \] The authors study the smoothness properties of the reduced energy \(\mathcal{F}_{\min }(t,\ell )=\min \{\mathcal{F}(t,\ell ,u):u\in \mathcal{U} (t,\ell )\}\), which allows to build the energy release rate \(\mathcal{G} (t,\ell )=\lim_{h\rightarrow 0^{+}}\frac{\mathcal{F}_{\min }(t,\ell )- \mathcal{F}_{\min }(t,\ell +h)}{h}\) and they compute the expression of \( \mathcal{G}(t,\ell )\) in terms of \(W\) and its gradient or in terms of the Eshelby tensor. The authors finally define the appropriate scaling in the context of small displacements than in the case of large domains, and the appropriate notion of \(\Gamma \)-convergence, which allows to study the convergence of energy release rates and of crack evolutions.