Robust equilibrated a posteriori error estimator for higher order finite element approximations to diffusion problems (Q2288367): Difference between revisions

The authors present an a posteriori error estimator for the numerical solution of \[ \left\{ \begin{array}{rcll} -\mbox{div}(A\nabla u) & = & f, & \mbox{in }\Omega\\ u & = & 0, & \mbox{on } \Gamma_D,\\ (-AA\nabla u)\cdot \mathbf{n} &= & g_n, & \mbox{on } \Gamma_N. \end{array} \right.\tag{1} \] Here \(\Omega\) is a bounded polygonal domain in \(\mathbb{R}^2\) or \(\mathbb{R}^3\) with boundary \(\partial \Omega =\Gamma_D\cup\Gamma_N\), \(\Gamma_D\cap\Gamma_N=\emptyset\). The function \(g_n\in L^2(\Gamma_N)\) is given and the matrix function \(A(x)\) is symmetric and positive definite on \(\Omega\) and the smallest eigenvalue is bounded away from zero. With these assumptions the weak formulation of (1) in \(H^1_D(\Omega):=\{v\in H^1(\Omega)\:|\: v|_{\Gamma_D}=0\}\) has a unique solution \(u^\ast\). To approximate \(u^\ast\) the authors use a continuous finite element space of order \(k\), \(k\geq 1\). In Sections 3 and 4, the authors describe an a posteriori error estimator that is based on the construction of an equilibrated flux and the Prager-Synge theorem. The calculation of the flux is given in Section 4. After formulating global and local error estimates in Section 5, it is shown in Section 6 that the proposed estimator is efficient. In the final Section 7, the authors present two numerical examples in two dimensions for \(\Omega=[-1,1]^2\) and for linear and quadratic elements. The a posteriori error estimates for both cases show the predicted behavior.