A posteriori error estimations of a coupled mixed and standard Galerkin method for second order operators (Q2469608)

The authors consider the discretization methods of the second-order boundary value problem: for \(f\in L^{2}(\Omega ),\) let \(\theta \in H_{0}^{2}(\Omega )\) be the unique solution of \(\text{div}(A\nabla \theta )=-f\) in \(\Omega ,\) where the matrix \(A\in L^{\infty }(\Omega ,\mathbb{R}^{2\times 2})\) is supposed to be symmetric and uniformly positive definite. They introduce some natural conditions on the finite element spaces and use the Clément interpolation operator. Since the meshes do not necessarily align at the interface, the resulting spaces are necessarily nonconforming. The reliability and the efficiency of the estimator are proved. Some numerical tests are presented and confirm the theoretical bounds.