A practical hybrid interior error estimator for localized h‐adaptive FEA

A practical method for localized h ‐adaptive error estimation is presented based on interior estimates of the Galerkin solution. A previously published hybrid interior error estimator is revisited here and proper bounds are established. It is shown that in the present form of the estimator both the local accelerated convergence and the global superconvergence properties are maintained. The estimator is based on energy norms and all the computations are based on groups of connected elements. The resulting form of the estimator is shown to be simpler and more amenable to computational implementation than the previous one. Two plane elasticity problems are chosen as examples and both structured and h ‐adaptive global initial meshes are considered to compute the convergence characteristics of the solution in a few preselected zones. The solutions are benchmarked against conventional global h ‐adaptive superconvergent error estimators.