Convergence of a space semi-discrete modified mass method for the dynamic Signorini problem (Q2379855)

The dynamic Signorini problem models infinitesimal deformations of a solid body which can come into contact with a rigid obstacle. The body occupies a reference domain \(\Omega \subset {\mathbb{R}}^d\) (\(d=2,3\)) during a time interval \([0,T]\). The boundary of the domain \(\Omega\) is piecewise smooth, so that its outward normal, \(n\), is well defined almost everywhere at the boundary. The material is linear viscoelastic (Kelvin-Voigt model). Then a precise formulation of the dynamic Signorini problem can be given by \[ \begin{aligned} &\rho \ddot{u}-\text{div} \sigma=f, \sigma = \mathcal{A}:\varepsilon + \mathcal{B} : \overline{\varepsilon}, \;\varepsilon = \frac{1}{2} (\nabla u + \nabla u^T)\,\, \text{ in }\,\, \Omega \times (0,T);\\ & u_n \leq 0, \;\sigma_{nn} \leq 0, \;\sigma_{nn} u_n = 0\,\, \text{ on \,\,} \Gamma^C \times (0,T);\\ &\sigma\cdot n =0 \,\,\text{ on } \Gamma^N \times (0,T), \;u=0 \,\, \text{ on }\,\, \Gamma^D \times (0,T);\\ & u(0) = u^o, \dot{u}(0) = v^o\,\, \text{ in }\,\, \Omega. \end{aligned} \tag{1} \] The authors consider the following variational formulation of equations (1): seek \(u \in L^2(0,T,K) \cap H^1 (0,T,V) \cap C^1 ([0,T],M)\) such that for all \(v \in L^2 (0,T,K) \cap H^1 (0,T,M)\), \[ \begin{aligned} \int^T_0\{-m(\dot{u},\dot{v} - \dot{u}) &+ a (u,v - u) + b (\dot{u},v -u) \} dt + m (\dot{u}(T), v(T) \\ &- u(T)) - m (\nu^0, v(0)-u^o) \geq \int ^T_0 l (t, v-u)dt. \end{aligned} \tag{2} \] The main result: the convergence, up to a subsequence, of the space semi-discrete solutions to a solution of the continuous dynamic Signorini problem in the case of viscoelastic material. The proof uses a priori estimates for space semi-discrete solutions and compactness arguments. The reviewer notes that the mass modification at the contact boundary requires a special care when passing to the limit as \(k \rightarrow \infty\). Here in both cases (penalty method or finite-dimensional variational inequality), the a priori estimate on the velocity delivered by the viscosity term plays a crucial role in the proposed proof (this argument cannot be extended to the vanishing viscosity limit).