Injectivity and self-contact in second-gradient nonlinear elasticity (Q2406078)

The authors prove an existence result for globally injective weak solutions to mixed boundary-value problems of second-gradient nonlinear elastostatics. They consider the energy \(E[f]=\int_{\Omega }[W(\nabla f,\nabla ^{2}f,x)-(b\cdot f+B\cdot \nabla f)]dV-\int_{\Gamma _{c}}[\tau \cdot f+\mu \cdot [ \nabla fn]]dS\), where \(W\) is coercive, continuously differentiable with respect to its two first variables and satisfies growth, material objectivity and polyconvexity with respect to its second variable properties. Here \(\Omega \) is an open and bounded subset of \(\mathbb{E}^{3}\) with continuously differentiable boundary \(\partial \Omega \). The deformation and its gradient are prescribed on a portion \(\Gamma \) of \( \partial \Omega \) by \(f_{0}\in W^{2,p}(\Omega ,\mathbb{E}^{3})\) and \(\Gamma _{c}=\partial \Omega \setminus \Gamma \). The authors define the set \( \mathcal{A}\) of admissible deformations which are locally invertible, orientation-preserving and globally injective as \(\mathcal{A}=\{f\in W^{2,p}(\Omega ,\mathbb{E}^{3})\mid f-f_{0}\in W_{\Gamma }^{2,p}(\Omega , \mathbb{E}^{3})\), \(\nabla f(x)\in GL^{+}(\mathbb{E}^{3})\) for all \(x\in \overline{\Omega }\), \(f\) is injective on \(\Omega \}\). The authors first prove the existence of a minimizer \(f^{\ast }\) of \(E\) on \(\mathcal{A}\). They here use classical functional analysis tools. The main result of the paper proves the existence of \(f^{\ast }\in \mathcal{A}\) and of a Radon measure \( \sigma \in C(\partial \Omega )^{\ast }\) such that \(\left\langle DE[f^{\ast }],h\right\rangle _{W^{2,p}(\Omega ,\mathbb{E}^{3})}+\left\langle \sigma ,h\cdot n_{f^{\ast }}\right\rangle =0\) for all \(h\in W_{\Gamma }^{2,p}(\Omega , \mathbb{E}^{3})\) and \(\sigma \) satisfies \(\left\langle \sigma ,z\right\rangle _{C(\partial \Omega )}=0\) for all \(z\in C(\partial \Omega )\) which satisfy \(z(x_{1})+z(x_{2})=0\) wherever \(f^{\ast }(x_{1})=f^{\ast }(x_{2})\) and \(x_{1}\neq x_{2}\). \(\sigma \) is non-negative and is supported on the coincidence set \(S_{f^{\ast }}\) of \(f^{\ast }\) defined as \(S_{f^{\ast }}=\{x\in \overline{\Omega }\mid \exists w\in \overline{\Omega }\setminus \{x\}\) such that \(f(x)=f(w)\}\). For the proof, the authors essentially use convexity tools, the compactness of \(S_{f^{\ast }}\) and a variational inequality associated to the energy \(E\).